1,300 years is long enough: it’s pens down for the exam hall
January 16, 2023
Liz Johnson, Helen Partridge & Phillip Dawson
In this post the use of the exam hall and its relationship to authentic assessment in a digital world is put to the test by Alfred Deakin Professor Liz Johnson, Deputy Vice Chancellor Academic Portfolio, Professor Helen Partridge, Pro-Vice Chancellor Teaching and Learning and Professor Phillip Dawson, Co-Director Centre for Research in Assessment and Digital Learning.
Lockdown responses to the COVID pandemic during 2020 and 2021 forced learning and assessment online. As universities move towards a new normal for teaching and assessment delivery, calls to revert to on-campus invigilated exam hall exams are appearing. Concerns focus on online assessment security with fears of a rising tide of cheating featuring in news media. The recent emergence of ChatGPT and other artificial intelligence tools has led to further calls to lurch back to the exam hall. The logic behind these calls is that we can’t really be sure students are doing the right thing if we are not watching them. But exam hall exams have their own problems – authenticity, assurance of learning and challenges for access. We need a serious conversation about exams and assessment to avoid the traps of the past.
Exam hall exams are an exercise in restriction and endurance. Students are given clear conditions: which materials they may use, when they must be in the building, where they must sit, who they can talk to (usually nobody), and how they must ask for permission if they need to use the bathroom (which for most students is the only reason allowed to get out of their chair). Then, students in the typical exam spend two to three hours writing their responses by hand. This is the way it is, and the way it has been for 1,300 years, dating back to some of the earliest recorded exams in China.
Assessment is ultimately about making judgements about what students are capable of. In making these judgements, context matters. The constraints of the exam hall are a poor representation of the sorts of contexts students will need to work in when they graduate. The context of the exam hall encourages learners to focus on exam technique and time management. It is hard to see how writing at the maximum rate with limited time for review under these constrained conditions prepares students for normal work or other high-stress environments where consultation and external oversight are likely to be much more important. Exam hall exams may carry the largest weighting for a subject/unit but are often the least authentic and inclusive assessment.
Given the constrained environment, exam hall exams ignore a large range of intended learning outcomes of study. All Australian universities require students to demonstrate a range of learning outcomes from their course. Typically, in addition to discipline knowledges and skills, these include communication, critical thinking, global citizenship, teamwork, independence, problem-solving and information literacy. Exam hall exams cannot provide evidence of achievement against most of these learning outcomes. Exam hall exams also fail to provide the crucial feedback that fosters learning. Students are not permitted to revise and improve and often have very limited access to find out what went ‘wrong’. Exam hall exams are lazy assessments – taking up considerable time and resources for limited value. In a crowded curriculum, can we really afford to expend precious assessment space on narrow tasks?
Two drivers for exam hall exams are apparent: concerns about cheating and challenges in delivering assessment at scale. Fears of increasing rates of cheating with online delivery and the apparent rise of contract cheating must be addressed. We know that cheating is considerably undetected even with the use of exam hall exams and that artificial intelligence poses new challenges. However, we should be focussing on contemporary answers to these challenges that encourage their best work from students. Resorting to exam hall exams won’t solve these challenges as more authentic assessment is required by students, industry and community partners and professional accreditation bodies.
There is no perfect singular solution to the problem of cheating, and instead we should be looking for layers of approaches that contribute to the security of our degrees at a program level. Exams have fostered a cottage industry of cheating since their earliest days, and they continue to be the target of industrious cheaters pedalling a range of custom hardware such as cheating calculators. Thinking that a switch back to exams solves the problem of cheating is not just naïve, it’s dangerous.
The challenge of providing equitable assessment at scale is considerable but the online environment can scale-up well. At Deakin University, we have moved our exam-hall exams into a range of online assessment tasks including more authentic tasks and online formats for end-of-semester assessment. Authentic online assessment tasks such as extended projects, simulations, case studies, group assignments can use the benefits of the online environment in gathering and curating information, analysing and presenting data, working asynchronously in groups and creating digital footprints of the process of developing submissions. We use online invigilation sparingly, usually where required by external accreditation, and note the challenges in using online proctoring systems.
In 2019, 686 of our subjects/units in Trimester/Semester 1 and 2 used an invigilated exam hall exam. In 2022, only 489 subjects/units during the same study period had an end of unit assessment, with only 112 requiring an online invigilated exam, a reduction of 84%. The remaining 377 subjects/units used open book assignments.
Noting the considerable effort to change between formats, we expect steady state delivery to be at least cost-neutral. We are already seeing a shift in focus from the process of managing exams to working on good assessment design.
When we assess our students, we should be judging and celebrating what they have achieved. In awarding a degree, we certify that our graduates have met all the learning outcomes of our degrees. There is decades of excellent research and practice to draw on to make assessment more valuable and more secure. In a digital world, exam hall exams are an anachronism that we can do without.
Excellent provocation. And while we are at it can we start challenging the supremacy of behavioural learning outcomes? These are only useful for such small range of learning, and not at all for much of the most important learning students need (and we seek to engender). Yet they have become the darling of uni admin as a way to make academics accountable…surely we have better ways to do this.
Pingback: Frequently asked questions about generative AI at Sydney – [email protected]
Pingback: CRADLE News: Congratulations to Dr Rebecca Awdry on her graduation!
Thank you for nice information.