This is part two of a pair of articles exploring the “exam reflex” facing schools and universities, and being driven by the news. In part one I looked at the AI cheating panic sucking up all the oxygen in mainstream media. I wrote about the “urgent threat” headlines, the Learning First report that is more nuanced than its coverage, and the knee-jerk reaction underneath it all: coursework is compromised, exams are safe, so let’s have more exams.
In part two, I’ll argue that exams-as-default-position is not only bad pedagogy, but also in direct opposition to what assessment authorities require of schools. Essentially, we cannot say “AI is a threat to uninvigilated assessment, and therefore we must do more exams” because our authorities do not permit that stance. Not only that, but it is demonstrably incorrect to say that “the students must do external exams at the end of the year, and therefore it makes sense to do their coursework under exam conditions.”
I went through the current policy handbooks, certificate manuals and syllabus documents of every senior secondary assessment authority in Australia: the QCAA (Queensland), NESA (NSW), the VCAA (Victoria), the SACE Board (South Australia, whose system also runs in the NT), SCSA (Western Australia), the ACT’s Board of Senior Secondary Studies, and Tasmania’s TASC. Each quote in this article is verbatim from an official document, which I’ve linked for your reading pleasure.
If you’re reading this from outside of Australia, I’d encourage you to go over your own jurisdiction’s assessment advice in just as much detail. I’d wager that you’ll find a few surprises of your own.
No Australian certificate is built on exams
Let’s kick off this deep dive by looking at the architecture of secondary school certification. The school-assessed share of a student’s final result is 50% in NSW and WA, up to 50% in Victoria (varying by subject), 75% in most Queensland General subjects (50% in maths and the sciences), a fixed 70% in every SACE Stage 2 subject — and 100% in the ACT, which has run senior secondary certification on continuous school-based assessment, in its own words, with “no external subject-based examinations”, since 1976.
The SACE Board states: “In a Stage 1 subject, all student work is assessed by the school. In a Stage 2 subject, 70% of student work is assessed by the school, with 30% assessed externally” (SACE Board). And that external 30% isn’t necessarily an exam either: “There are three types of external assessment: investigations, performances, and examinations” (SACE Board). In SACE English, the externally assessed component is a comparative analysis: an independent study task, marked outside the school, with no exam room in sight.
Tasmania’s TASC, in its student-facing explainer, states: “Even in courses that include external assessment, the largest proportion of the overall award is from the internal assessment by the classroom teacher.” Whole categories of course carry no external assessment at all, such as Queensland’s Applied subjects, WA’s General and Foundation courses, Tasmania’s Level 1 and 2 courses, all of SACE Stage 1, and everything in the ACT (where ATAR scores rely on the additional ACT Scaling Test or AST).
If exams were the system’s true preference and coursework the concession, you would expect the weightings to lean the other way.
The authorities say, frequently, that exams can’t measure everything
This is the part where colleagues at VCAA, NESA, QCAA and so on often get frustrated: there is a misconception that exams are the be-all and end-all. This is the problematic narrative driven by a media that publishes study score rankings and NAPLAN league tables. It is the parent pressure on schools, and sometimes the perspective of school leadership teams. Sometimes, it is a misconception of how the exams and senior school assessment systems work. But it is not evidenced in the assessment authority documents.
The QCAA’s QCE and QCIA policy and procedures handbook — the foundational document of Queensland’s senior assessment system — says this:
“External assessment is not privileged over summative internal assessment. It is a mechanism for adding evidence that is different, yet equally valuable, to determine a student’s overall achievement.”
– Section 10
Not privileged. The same handbook grounds the whole design in a validity principle that directly contradicts the exam discourse:
“The validity of assessment is improved by assembling evidence of student achievement from a variety of assessment techniques and conditions. Reliability of assessment is improved by providing students with multiple opportunities to demonstrate their knowledge, understanding and skills, as well as by collecting evidence at different times and under different conditions.”
– Section 1.3
Pushing every task into the same timed, supervised, unseen conditions doesn’t make assessment more rigorous on this view; it makes it less valid.
NESA – the authority responsible for the most exam-infamous credential in the country – tells students directly: “School-based assessment tasks can measure a wider range of skills and knowledge than an external exam can.” Its curriculum guidance says formal school-based assessment exists to gather evidence “in different ways to the HSC examinations”. The ACT’s board justifies its examless system in stronger terms still: school-based design gives the teacher “flexibility to design assessment of different types that take into account student interest, background knowledge and experience, leading to more student engagement and more validity of assessment results” (BSSS).
And the SACE Board, in its current strategic direction, is moving away from test-shaped evidence right now:
“While ensuring students continue to have fair, valid and reliable results, the SACE Board will seek a wider range of evidence to more accurately show what students know, understand and can do. Teachers will be encouraged to submit evidence of learning that is less formal – things other than essays and test results, for example – that are nevertheless legitimate and can count towards a student’s attainment of the SACE.” (SACE Board)
Where the documents do assign exams a purpose, it is consistent and narrow: a common, standardised reference point. NESA’s explanation of moderation is direct about why the exam anchors the system: “The only task that all students do that is the same, and is marked in exactly the same way is the HSC exam.”
WA’s WACE Manual lists the purposes of its ATAR exams as results that “complement and support school assessment”, diagnostic feedback for teachers, statistical moderation, public confidence, university ranking. Nowhere in any of these documents does any authority claim that exam conditions are the best general way to find out what a student knows.
Assuming most people reading this post are educators, we already know that exams aren’t the best way to judge knowledge and skills. That doesn’t help the framing in the media, though. And unfortunately, there is still the persistent misconception that “QCAA (or whomever) values the exam more highly…”. So let’s address that, too.
The rules don’t just permit non-exam coursework: they often mandate it and cap exams
In Queensland, every General syllabus prescribes the technique of each internal assessment, and examination is one technique among six: “projects, investigations, extended responses, performances, products and examinations”. Maths students sit two school exams; science students sit a 60-minute data test. But Visual Art, Drama and Digital Solutions have zero examination-technique internal assessments: their entire 75% internal component is investigations, projects and performances, and a school cannot replace them with tests. Even where Queensland does prescribe an internal “examination”, it sometimes deliberately removes the surprise: in English and Modern History, QCAA’s assessment materials describe the task or sources being provided to students a week ahead. Timed and supervised, yes, but unseen, deliberately not.
In WA – the only state that mandates a school examination in every ATAR course – the same syllabus tables that require the exam also cap it. English ATAR Year 12: Responding 35%, Creating 35%, Examination 30%. Seventy per cent of the school mark cannot be the exam, by regulation. Physics allows more (tests 30% plus exams 50%) but still walls off 20% for science inquiry work: “There must be at least one experiment, one investigation and one evaluation and analysis completed in this pair of units” (Physics ATAR Year 12 syllabus). The WACE Manual is explicit that “The weighting for each assessment type must conform to the weighting (or weighting range for some courses in Year 11) specified in the assessment table in the relevant syllabus”.
In NSW, the structure many people half-remember – four tasks, one formal written exam capped at 30% – comes from NESA’s own published assessment programs: “4 assessment tasks, including: a minimum weighting for an individual task of 10%; a maximum weighting for an individual task of 40%; one task that is a formal written exam with a maximum weighting of 30%” (English Advanced Stage 6). One exam-condition task out of four, capped at 30% of the school mark is NESA’s model of good assessment design.
However, since an October 2024 rule change, that model is suggested rather than binding: ACE Rule 2.1.2 now says schools “may follow the assessment program provided by NESA, but have the authority to determine the number, type and weighting of assessment tasks”. A NSW school could now legitimately go exam-heavy, but it would be doing so against NESA’s own stated purpose for school-based assessment, using a flexibility that was introduced to let schools innovate, not to let them retreat.
In South Australia, the SACE Board runs a formal two-level taxonomy where tasks are either “directly supervised” (teacher present at all times) or “indirectly supervised” (periodic checks, drafts, work-in-progress), and subject outlines decide which subjects need any directly supervised coursework at all. In Mathematics, all skills and applications tasks are supervised. In Biology and Chemistry, there are at least two supervised tasks, capped at 90 minutes each. English and Modern History have none required, anywhere in the 70% school component. Even SACE’s supervised maths tasks permit “Electronic technology and up to one A4 sheet of paper of handwritten notes”, and from 2025 the Board removed its last remaining requirement that one task be done with no calculator and no notes. In the middle of the AI panic, an Australian assessment authority abolished its last closed-book coursework rule.
Moderation systems protect teacher judgement from the exam
The loudest version of the exam discourse says fine, but the exam is what keeps everyone honest. Coursework only counts because exams provide the accountability and authentication. The assessment and moderation documents say otherwise.
In the three states that statistically moderate school marks against exam performance, the authorities are at pains to explain that the exam adjusts the scale of a school’s marks, never the teacher’s ranking of students. I remember this from my own 7+ years as a VCE English exam assessor, where every year I attended the initial benchmarking and moderation meetings.
The VCAA states that “Statistical moderation does not change the rank order of students as determined by the school’s school-based scores. A student given the top score for school-based assessments by his or her school will have the top score after statistical moderation, no matter how they perform on the exam(s)” (VCAA). NESA also states that “Each student’s rank stays the same when school assessments are moderated” (NESA). WA’s version says the same thing in its combined-scores explainer.
And four jurisdictions don’t use the exam to moderate at all. Queensland’s “External assessment is included in all General subjects, but is not used to scale a student’s internal assessment result.” Quality assurance there is expert review, with every assessment instrument endorsed before use, and samples of marked student work confirmed afterwards. SACE moderates by having trained moderators read actual student work against published performance standards. Tasmania runs peer-review meetings where teachers from different schools assess the same student work. The ACT does structured, consensus-based peer review at system-wide moderation days. Comparability, in more than half the country, is a function of professional judgment; not of an exam.
Defying the authorities
So here is where the exam discourse actually lands. A school that responds to AI by converting its assessment program to wall-to-wall timed, invigilated, unseen conditions would breach the mandatory syllabus assessment tables in Queensland and Western Australia, exceed what SACE subject outlines permit outside maths and science, contradict most VCE study designs, and defy the published design principles of every assessment authority in the country: the bodies that, between them, certify every Year 12 student in Australia.
I’ve been banging on about this argument since before GenAI existed, so I’ll clamber down from my high horse at this point. But the quotes above aren’t mine. They belong to the authorities, and they describe a system designed so that most assessment is not an exam, because most of what we want to know about a student cannot be examined.
The obvious counter-argument, and the one hinted at in the media, is that these documents were (mostly) published before the release of ChatGPT. My position: AI didn’t change the nature of academic integrity, and it didn’t change the values and positions that underpin good assessment practice.
The AI cheating crisis is an authentication crisis. Schools already own the skills, knowledge, and professional expertise for authentication, and the discourse is at least right that strengthening it is urgent, system-level work. But if we let the panic talk us into the exam hall instead, we won’t be restoring the system’s integrity. We’ll be dismantling its design, in defiance of the same authorities that award the certificates.
I’ve covered a lot of ground in this article. If you have suggestions, corrections, or questions, please don’t hesitate to let me know!
Want to learn more about GenAI professional development and advisory services, or just have questions or comments? Get in touch:
Leave a Reply