Curriculum accuracy — and how to check us

Three different things, measured separately

Most platforms publish one accuracy figure. One figure hides the third thing on this list, which is the one that matters most if a reviewer starts asking questions. So we keep them apart.

The number we found when we marked our own work

On 2 August 2026 we drew a random sample of 100 of our 812 outcomes, stratified by learning area, and put each one next to ACARA's official wording. Our director marked every one: does our description actually describe this code?

He disagreed with our tagging on 13 of the 100. Allowing for the sample size, our mis-tagging rate across the whole program is at most 19.5% (95% confidence, one-sided Wilson bound). Across the whole program that is somewhere around a hundred descriptions that need work.

That is the figure as we found it, and it is the figure we are publishing — not a corrected one. Six of the thirteen were genuine errors and we fixed them the same afternoon, but we have not gone back and re-scored the sample to make the headline smaller, because a number you only publish once it flatters you is not a measurement. The sample, the seed and the tool are named at the bottom of this page so the same draw can be repeated against us.

What the 13 actually were. Not one error repeated. Four were content filed under a neighbouring code — we described the rock cycle under the plate-tectonics descriptor when v9 gives the rock cycle its own code, which we then never claimed. Two were a straight swap, where content we deliberately teach a year early kept the code belonging to the slot it moved into rather than the code that describes it. Two were us teaching more than the descriptor asks and claiming the descriptor anyway. Five were honest differences of emphasis, where another teacher might well have marked them the other way.

The first six were errors and were corrected the same day. Every one of them increased what we can honestly claim rather than reducing it: the teaching was already happening, and the correction was to file it under the code that describes it. That is why the count above went from 812 outcomes to 818 — six descriptors we had been earning and never claimed.

We are not re-marking this sample to show a better number. Once a reviewer has seen the questions, they are no longer measuring the same thing. The next figure will come from a fresh draw under a new seed, and it will be published beside this one whether it is better or not.

What we found when we checked ourselves

In July 2026 we audited all 771 outcome codes across 85 units against ACARA's published content descriptions. The first pass used a reference file that had been truncated at 150 characters — so roughly a third of the official descriptions were cut off, often mid-sentence, and the review was judging our work against incomplete text.

We rebuilt the reference in full and ran a second, independent pass over every learning area. It corrected 20 curriculum codes and 81 outcome descriptions. The worst single case: a Year 9 mathematics outcome whose description said the opposite of the curriculum, because the truncation fell one character before the clause that defined its scope.

We are telling you this because it is the kind of thing that normally gets fixed quietly. The lesson we took from it is that a check is only as good as what it checks against — which is why the reference data and the audit tool now both live in our codebase rather than in someone's head.

Where we teach something in a different year to v9

This is deliberate, and it is the part we most want you to see before a reviewer does.

We build integrated units — a meaningful whole that a child works through, rather than disconnected subject blocks. When you do that, the adventure decides what is needed next, not the filing sequence. A story that needs someone to speak needs quotation marks, whether or not the child has reached the year v9 introduces them. Meeting an idea early and again later is good teaching; the Australian Curriculum itself is built to be revisited.

But it does mean that if you generate a plan straight from our codes, these are the rows that will not line up with v9's year sequence. Here is every one of them, with the code to cite if you are asked.

Two of these are not integration choices but genuine gaps in v9 itself: Japanese has no Years 9–10 descriptor for Japan's own language variation, and Foundation Technologies has no descriptor for sequencing steps. We teach both anyway, and have said which code we file them under.

What we don't claim

• We don't claim the 13% is independent review. It was marked by our own director, who set the standard those descriptions were written to. A self-audit is worth publishing — it is how we found the errors — but it shares our blind spots by construction, and the honest expectation is that an outside reviewer would find more, not fewer. The independent figure will be published when qualified teachers who have not seen our reasoning have marked a sample. That review is built and we are recruiting for it now; until it reports, this page carries a first-party number and says so.
• We don't claim to be right where teachers disagree. Two qualified people tagging the same lesson against v9 will not always agree — the descriptors are written loosely enough to be argued over in good faith. Five of our own director's 13 disagreements were exactly that kind: a difference of emphasis, not a filing error. So the 13% above is one marker's read, and it is the ceiling rather than the floor of what we would call wrong. The independent figure, when the teacher panel reports, will be shown as an agreement rate next to how often the reviewers agree with each other — because a disagreement rate means nothing until you know how much two experts disagree anyway.
• Years 11 and 12 are not measured here. Senior schooling in Queensland follows QCAA syllabuses, not Australian Curriculum v9 codes. Those units cite syllabus units instead, and this page does not cover them.
• This is not the registration decision. Alignment is our job; approving your programme is the Home Education Unit's. Always confirm codes at v9.australiancurriculum.edu.au before you submit.

How to check us yourself

The audit tool ships in our codebase. It draws a random sample using a fixed seed, so anyone can re-draw the identical sample and mark it independently — you do not have to take our word for which outcomes we looked at. The 13% above came from curriculum-audit.mjs sample 100 20260801, stratified by learning area. That draw was made against the 812 outcomes we published on 2 August, before the six corrections described above took the count to 818 — so running it at that commit hands you the identical hundred outcomes, and you can mark them yourself without reference to how we marked them.

It also refuses to produce an accuracy figure from automated text matching, and says why in its own source: when we tried that, it flagged 29 Arts outcomes of which one was a real error, and scored Mathematics perfectly clean while half that area needed correcting. An automated matcher is a hint, not a measurement.

© Australian Curriculum, Assessment and Reporting Authority (ACARA) 2010 to present, unless otherwise indicated. This material was downloaded from the Australian Curriculum website (accessed 1 August 2026). The curriculum codes and content descriptions we check against are unmodified; the unit summaries you read on this site are our own paraphrases and are therefore modified material. The ACARA material is licensed under CC BY 4.0.

ACARA does not endorse any product that uses the Australian Curriculum and makes no representations as to the quality of any such product. Any product that uses material published on the Australian Curriculum website should not be taken to be affiliated with ACARA or to have the sponsorship or approval of ACARA. Any errors in how we have applied the curriculum are ours.

Questions

Email admin@parksydeltd.com or call 0467 937 659. If you are preparing a registration submission and want the alignment for a particular unit explained, ask — that is exactly what this is for. See also our parent guide and the full curriculum map.