Grammar and punctuation is, by a clear margin, the weakest literacy domain Australian students sit in secondary school. In the 2025 NAPLAN cycle only 56.4% of Year 9 students reached 'Strong' proficiency or above in grammar and punctuation, the lowest figure of any domain tested that year, well behind spelling's 71.9% [WRI-07]. That weakness is not a snapshot; it is a decade-long trend. AERO's analysis of ten years of NAPLAN writing data found that for secondary students specifically — Years 7 and 9 — the criteria that deteriorated over time were Punctuation, Sentence Structure, and Vocabulary, a genuinely different pattern from the primary years, where the decline concentrated in Text Structure and Persuasive Devices instead [WRI-03]. Spelling, tellingly, was the only criterion in that whole study that improved [WRI-03] — a distinction that matters, because it means 'conventions' is not one uniform problem. Something is going right in spelling and going wrong in the sentence-level mechanics of grammar and punctuation, at exactly the age Argo's students are sitting scholarship and selective-entry exams.
The picture at Year 3 shows the same shape lower down the age range. Grammar and punctuation was the weakest of all five NAPLAN domains for that cohort too in 2025, with 81.1% of students reaching 'Developing' proficiency or above against a high of 92.4% in writing [LAN-02]. And this is not a domain in visible motion, up or down: the Australian Bureau of Statistics's own wellbeing dashboard shows Year 3 grammar-and-punctuation scale scores essentially flat across the first three post-reset NAPLAN cycles — 410.8, then 408.8, then 408.4 — with spelling similarly flat over the same period [LAN-03]. So the honest description is not a domain in crisis or a domain recovering; it is a domain stuck low, consistently, at both ends of the school system.
It is worth pausing on how that low result gets reported, because this register's own verification process caught a case of the story running ahead of the data. The Guardian's coverage of the 2025 results claimed 39.2% of students performed 'below expectations' in grammar and punctuation nationally, and glossed this as students 'struggling to recognise verbs and pronouns in sentences' [LAN-01]. On direct verification, that specific 39.2% aggregate could not be located anywhere in ACARA's own National Results Commentary, which reports only per-year-level breakdowns, and the 'verbs and pronouns' framing turned out to be the journalist's own editorialising, not a quote attributed to ACARA at all [LAN-01]. The figure was downgraded accordingly rather than repeated as an official statistic. That correction is itself part of the evidence here: even a widely reported, plausible-sounding number needs to be traced to its primary source before it does any work in an argument, and grammar and punctuation — being the domain most prone to alarming, simplified headlines — is a live example of why.
There is also a gender pattern specific to this domain that deserves its own weight. A longitudinal study of NAPLAN spelling and grammar-and-punctuation results from 2008 to 2021 found girls outperformed boys at every tested year level, in every year of testing, across the full fourteen-year span [LAN-09]. The grammar-and-punctuation gap was wider than the spelling gap at every point, and it widened with age — from roughly half a year of equivalent learning at Year 3 to well over a year by Year 7 [LAN-09]. More strikingly, while the spelling gender gap closed in more recent testing years at every level, the grammar-and-punctuation gap only closed at Years 3 and 5; at Years 7 and 9 it widened instead [LAN-09]. Boy cohorts made two years of equivalent progress between tests only once across the entire study, on grammar and punctuation, and fell behind fastest between Years 5 and 7 [LAN-09] — precisely the transition years where Argo's scholarship-preparation work concentrates.
So the weakness is real, persistent, and worse for boys as they move into secondary school. The paradox is what the intervention evidence says about fixing it. The same Writing Next meta-analysis that identifies effective writing interventions found that grammar instruction produced a small but statistically significant negative effect on writing quality, both for the general student population and for low-achieving students specifically [WRI-14]. The meta-analysis's own synopsis does not spell out exactly what counted as 'grammar instruction' in the studies it pooled, but the wider Writing Next literature describes this finding as applying to grammar taught as a subject in its own right — parsing sentences, naming parts of speech — rather than grammar knowledge applied directly to a student's own writing. On that reading, drilling grammar in isolation does not just fail to help; it measurably hurts. What does work, in the same meta-analysis, is sentence combining: teaching students to build more complex sentences by joining simpler ones produced a moderate, statistically significant effect on writing quality (effect size 0.50), with an almost identical effect (0.46) for low-achieving students specifically [WRI-12]. The lesson is not that grammar knowledge is irrelevant — it is that the same content, taught as a rule to memorise versus a tool to build sentences with, appears to produce opposite results.
Two further findings round out what actually moves this domain. Spelling — the one criterion in the AERO decade study that improved rather than declined [WRI-03] — has its own strong, direct evidence base: a meta-analysis of 53 studies and over 6,000 students found formal spelling instruction produced a moderate-to-large effect on spelling accuracy (ES = 0.54 against no instruction, 0.43 against informal approaches) and, notably, a large effect on generalisation into students' own writing (ES = 0.94) [LAN-06]. That generalisation figure is the one worth underlining: spelling taught formally does not stay confined to a spelling test, it shows up in the student's actual prose. Morphological instruction — teaching prefixes, suffixes and root words — carries a related benefit across literacy outcomes generally, with particular benefit for less able readers [VOC-07], giving students a transferable strategy for both spelling and unfamiliar vocabulary rather than a fixed list to memorise.
Put together, the practice implication is specific rather than vague. Conventions should not be taught as a standalone drill divorced from a student's own writing — the evidence says that approach can actively backfire [WRI-14]. It should be taught where the sentence-level and vocabulary declines actually live: inside the writing itself, through techniques like sentence combining that build grammatical control by using it [WRI-12], reinforced by the kind of formal, structured spelling and morphology instruction shown to generalise into real prose [LAN-06] [VOC-07]. That is where accuracy and quality meet — not in a worksheet of underlined verbs, but in a sentence a student has just built, made stronger, and can now build again.
