Section 2 · Written Communication
GAMSAT Essay Marking Criteria
Two criteria, three independent scorings, and no published weighting between them. What follows is what ACER actually says, separated from what the preparation market has invented around it.
- 2official criteria
- 3independent scorings
- Nonepublished weighting
- 0marks lost for spelling
The short version
You are marked on thought and content and on organisation and expression. Every response is read and scored 3 times independently. ACER has never published a weighting between the two criteria, and the widely quoted 60/40 split is a preparation provider’s estimate, not an ACER figure. Your opinion is never marked wrong. A memorised essay is.
The two criteria, in ACER's own words
ACER prints the criteria in a boxed panel in the Information Booklet. There are two of them, each defined by two sub-points, and they are reproduced here in full because paraphrase is where most of the misinformation on this topic starts.
| Criterion | ACER’s gloss | What it covers |
|---|---|---|
| Thought and content | the quality of what is said |
|
| Organisation and expression | the quality of the structure developed and language used |
|
Note what is not in that panel: no word count, no paragraph count, no essay genre, no mark scheme, no band descriptors. ACER assesses two things and declines to tell you how it converts them into a number.
Why there is no 60/40 split
Search for how Section 2 is marked and you will repeatedly find that thought and content is worth 60% and organisation and expression 40%. That figure does not come from ACER.It comes from a preparation provider’s marking sheet — and that same sheet states plainly that ACER has never revealed its exact marking criteria. A different provider says the two are weighted equally. Both cannot be right, and neither is sourced.
What ACER publishes instead is a statement of dependency:
ACER, on how language is assessed
“Control of language (grammatical structure and expression) is an integral component of a good piece of writing but it is not assessed in isolation. It is assessed insofar as it contributes to the overall effectiveness of the response to the task.”
That is more useful than a percentage, and it points the opposite way to how most candidates prepare. Language is not a second bucket of marks you can fill to compensate for thin ideas. It is assessed as the vehicle of thought. Elegant prose carrying a shallow argument has nothing to be effective at, so there is no separate pot of marks for it to win.
The practical consequence: time spent making a mediocre idea sound better is worth less than time spent making it a better idea. If you are choosing between polishing a sentence and developing a paragraph, develop the paragraph.
The five questions your markers ask
ACER’s preparation guidance lists the questions “considered by the examiners”. This is the closest thing to a rubric that exists, and it is public.
| # | Question examiners consider | Criterion |
|---|---|---|
| 1 | What kinds of ideas have you developed in response to the prompt? | Thought and content |
| 2 | To what extent and how well have you explored the ideas and issues in the prompt? | Thought and content |
| 3 | Have you developed a structured and organised piece of writing? | Organisation and expression |
| 4 | Have you written clearly and fluently? | Organisation and expression |
| 5 | Is the language you used precise and appropriate for the topic? | Thought and content / expression |
Question 2 is the one worth rereading. It asks to what extent and how well you have explored the ideas — not how many ideas you produced. That rewards depth and punishes the common instinct to sweep through four examples in four paragraphs. Two ideas taken somewhere are worth more than five listed.
What ACER explicitly penalises
Three things are named in the Information Booklet as scoring badly. All three are worth knowing before you buy any preparation product.
| Penalised | What ACER says | Consequence |
|---|---|---|
| Off-topic responses | “Responses that do not relate to the topic will receive a low score.” | Low score |
| Pre-prepared essays | “Pre-prepared responses… will receive a low score.” “Memorised essays generally score poorly.” | Low score |
| Borrowed or AI-generated text | “All responses will undergo plagiarism checks, including the use of Artificial Intelligence (AI)… including checking against previously submitted GAMSAT responses.” | Low score and possible misconduct |
The memorised-paragraph trap
ACER names preparation providers as a source of contamination directly: “Essays that borrow text or statements from pre-prepared essays (often provided by preparation providers) could be flagged as plagiarised even if you have not seen the original source. In these cases you could be plagiarizing without even knowing it.”
Read that carefully. If a provider sells model paragraphs to thousands of candidates and you write something close to one independently, the checker does not know which of you came first. That is why we do not sell model paragraphs and argue against buying them.
The reverse also holds, and it is reassuring: “Test takers are not assessed on the ‘correctness’ of the ideas or attitudes they display.” There is no house view to guess at. You cannot lose marks for the position you take, only for how thinly you take it.
What markers appear to reward — and where sources disagree
Everything above is ACER’s. What follows is not: it is drawn from preparation providers and from candidates reporting their own results, and it is labelled as such because it has a different evidential status. We include it because it is the only granular guidance in existence, and we flag the one place it openly contradicts itself.
| Pattern | Reported by | Status |
|---|---|---|
| Depth over breadth — develop fewer ideas further | Preparation providers and high scorers, in agreement | Consistent with ACER’s examiner question 2 |
| Nuance beats a clean side — grey areas where you can neither fully agree nor disagree | Candidates reporting 70–80+ | Candidate-reported |
| Avoid extremes; write as if your marker holds the opposing view | Candidate reporting 100th percentile | Candidate-reported |
| State your contention early and restate it at the close | Candidates reporting 80+ | Candidate-reported |
| Avoid cliché exemplars (Hitler, the World Wars) — markers are comparing you to every similar essay | Preparation provider | Provider guidance |
| Avoid reaching for vocabulary you do not control | Candidate-reported | Candidate-reported |
An open disagreement we are not going to resolve for you
On how much of a paragraph an example should occupy, the sources directly conflict. A tutor reporting 90 in Section 2 says examples “should make up no more than 1/2 of 1 sentence of each paragraph” because the section assesses the quality of thinking. A major preparation provider says the opposite: if your example is only one sentence, “that usually means the example is not detailed enough.”
Both are experienced sources. We report the disagreement rather than picking the one that sounds more confident, because there is no ACER statement that settles it.
Score bands — treat these as estimates
ACER publishes no Section 2 band descriptors and no target scores. The figures below circulate widely and are worth knowing as rough orientation, but they are preparation providers’ estimates of the distribution, not measurements of it.
| Reported figure | Source type |
|---|---|
| Around 59 is roughly the 50th percentile | Preparation provider estimate |
| Around 64 is roughly the 75th percentile | Preparation provider estimate |
| Above 70 is considered excellent | Preparation provider estimate |
| The average sits somewhere between 59 and 65 | Preparation provider estimate |
For what these numbers mean in an application, the useful page is what counts as a good GAMSAT score— and note that Section 2 is one quarter of your overall score under ACER’s reporting, but not every university weights it that way.
Common questions
How many people mark my GAMSAT essay?
Each response is scored independently 3 times. ACER states this directly in the Information Booklet. It is one of the details most preparation sites omit, and it matters: an unusual response is unlikely to be sunk by one marker having an off day.
Is thought and content really worth 60%?
There is no published weighting. The 60/40 figure comes from a preparation provider’s marking sheet whose own author states that ACER has never revealed its criteria. What ACER does say is that language is not assessed in isolation — it is assessed through its contribution to the response as a whole.
Can I lose marks for the opinion I argue?
No. ACER states that test takers are not assessed on the correctness of the ideas or attitudes they display. You are assessed on what you make of the task, not on whether a marker agrees with you.
Does spelling or punctuation cost me marks?
Not directly. ACER states that spelling and punctuation errors will not impact your scores, and the writing application has no spellcheck or autocorrect. Grammar and expression are assessed, but only through their effect on how clearly your ideas land.
Will ACER know if I used AI?
ACER states that all responses undergo plagiarism checks including for AI use, and that responses are checked against previously submitted GAMSAT responses. Beyond the detection question, memorised and generated material scores poorly on its own terms, because the criteria reward what you make of the specific prompt in front of you.