How reading is learned, and where Australian reading results stand. 24 verified findings — quoted verbatim, fully cited, independently checked (register v1.56.2, verified 2026-09-16).
Reading in Australia: Where We Stand and What Actually Helps
Start with the number that frames everything else: Grattan Institute estimates that roughly a third of Australian children cannot read proficiently — about eight students in a typical classroom of twenty-four [RDG-01]. Grattan puts a dollar figure on what that means over time, estimating the lifetime economic cost to Australia of today's poor reading outcomes, for the students hardest hit, at $40 billion [RDG-02]. It's worth being precise about what this number is: it's Grattan's own synthesis of the wider Australian evidence base (NAPLAN, PIRLS and related data), not a single primary measurement, and it underwrites Grattan's policy pitch — a national 'Reading Guarantee' committing to at least 90% proficiency, built around a structured-literacy approach and a Year 1 Phonics Screening Check [RDG-03]. That's advocacy, not a finding, but it's advocacy grounded in a real and measurable gap.
The government data mostly bears this out, though the picture is more stable than declining in the recent past. In PIRLS 2021, Australian Year 4 students scored 540 on average — statistically similar to 2016 (544) and higher than 2011 (527), with about 80% reaching the intermediate benchmark and Australia sitting mid-pack internationally (outperformed by six countries, similar to eight, ahead of twenty-eight) [RDG-10]. There's a persistent gender gap inside that average: 84% of girls met the benchmark against 77% of boys [RDG-11]. Zoom out to PISA and the longer trend is less comfortable. Australian 15-year-olds' reading scores fell by more than 25 points between the early 2000s and 2022 — a fall equivalent to roughly a year of schooling, meaning today's 15-year-olds read at the level 14-year-olds did two decades ago — though performance was flat between 2015 and 2022 (worth flagging: the OECD reports this 25-point figure for maths and reading combined, not reading alone) [RDG-12]. In PISA 2022, 79% of Australian students cleared the baseline (Level 2) reading standard against an OECD average of 74%, and 12% were top performers against 7% [RDG-13]. The PISA 2025 country note, published in September 2026, gives 77% against 69% and 11% against 6%. Australia's shares fell by two points and one; the OECD average fell by five and one, so Australia's margin over the average widened even as its own share fell. Reading was a minor domain in the 2025 cycle and the note prints no significance test for the change, so this is a newer figure, not a measured decline [RDG-21].
Where the data gets genuinely uncomfortable is equity. NAPLAN 2024 shows an average 67.0% of Australian students nationally achieving 'Exceeding' or 'Strong' reading proficiency, dipping to 63.0% by Year 9, with 10.3% needing additional support on average [RDG-14]. Underneath that headline sits a stark gap: roughly one in three Indigenous students (33.7%) fall into the 'needs additional support' category, against fewer than one in ten non-Indigenous students (8.6%); and only 24.0% of students in very remote schools reach 'Strong' or 'Exceeding', against 70.7% in major cities [RDG-15]. ACER's own PIRLS reporting confirms these socioeconomic, location and First Nations gaps have not closed since 2011 or 2016, even if it doesn't publish the exact comparative scores [RDG-11]. Whatever else is true about Australian reading, it is not evenly distributed.
On instruction, the strongest evidence base is for systematic phonics — with real nuance attached. The US National Reading Panel (2000) concluded that systematic, explicit phonemic awareness instruction directly improves reading and spelling, and recommended it for pre-school children and beyond who hadn't yet mastered phoneme concepts [RDG-04]; it separately found phonics instruction produces significant benefits from kindergarten through Year 6, with the strongest effects from systematic approaches [RDG-05]. A later meta-analysis of the Panel's own dataset (38 studies, 66 treatment-control comparisons) found a moderate overall effect, with early instruction (Kindergarten/Year 1) roughly twice as effective (d = 0.55) as instruction started after Year 1 (d = 0.27) [RDG-06]. We're flagging an honest discrepancy here: the commonly cited effect size for this analysis is d = 0.41, but our verification against the primary NRP report itself repeatedly finds d = 0.44 stated as the overall figure, and the 0.41 figure could not be independently confirmed — so we report it as 'approximately d = 0.4–0.44' rather than pretending to false precision [RDG-06]. The Education Endowment Foundation, working from a broader and more recent United Kingdom evidence base, rates phonics as worth about five months' additional progress on average, with particular benefit for disadvantaged pupils [RDG-07]. All of this evidence is US- or UK-sourced and skews to primary-age children — none of it is Australian, and 'early is better' is the one finding that repeats cleanly across sources.
Comprehension is a knowledge problem as well as a decoding problem, though. In the classic 'baseball study', prior topic knowledge outweighed general reading ability: weak readers who knew a lot about baseball comprehended and recalled a baseball text as well as strong readers who didn't, with prior knowledge affecting every recall measure and no interaction with reading ability [RDG-08]. That's one narrow, much-replicated 1988 US study of 64 students, not a general law, but it points at something real — comprehension leans on what you already know as well as on how well you can sound out or parse a sentence. Vocabulary and comprehension then reinforce each other: children with stronger vocabulary comprehend better, and their comprehension grows faster over time, partly because written text is a richer source of new vocabulary than everyday speech [VOC-10]. Nagy and Anderson's classic corpus study of United States printed school English estimated some 88,500 distinct word families exist in it and argued direct instruction alone could never cover that ground — implying most vocabulary growth has to come from wide reading itself [RDG-18].
That last point connects to the sharpest finding in this whole domain: reading volume compounds. In United States research, Cunningham and Stanovich found the average child at the 90th percentile of reading volume reads almost two million words a year outside school — more than 200 times the roughly 8,000 words read by a child at the 10th percentile — meaning a full year of the low-volume child's outside-school reading equals about two days for the high-volume child [RDG-09]. This is the empirical seed of the 'Matthew effect': early, small differences in reading skill and enjoyment compound into enormous exposure gaps over years. There's supporting evidence that the gap shows up in real outcomes: a UK birth-cohort study found children who read for pleasure frequently at ages 10 and 16 gained significantly more in vocabulary, spelling and maths by 16 than infrequent readers — a linked effect roughly four times the size of having a parent with a university degree, after controlling for prior attainment and background [LIT-04]. A follow-up on that same United Kingdom cohort found frequent childhood readers still had noticeably higher vocabulary scores at age 42, thirty years on, even after controlling for social background [LIT-09]. PISA 2018 data across OECD countries found students who read books often in print scored 49 points higher than infrequent readers, after accounting for socioeconomic background and gender [LIT-05]. All of this is correlational or carefully controlled observational data, not experiment — the researchers themselves use 'linked to', not 'causes' — but the consistency across independent, decades-apart studies is hard to wave away.
And reading for enjoyment specifically seems to matter more than reading time alone. OECD PISA analysis found students who read daily for enjoyment scored the equivalent of about a year and a half of schooling better than those who didn't — though the share of students reading daily for enjoyment fell across the OECD from 69% to 64% between 2000 and 2009 [RDG-16], and that decline is uneven: advantaged 15-year-olds were far more likely to read daily for enjoyment (72%) than disadvantaged students (56%), with Australia named as one of ten OECD countries where that gap exceeds 20 percentage points [RDG-17]. Locally, a large 2022–23 survey of roughly 12,000–13,000 Australian secondary students found 54% read more than once a week, fiction the most popular genre and print the preferred format — but also that 29% of students reported no reading engagement at all in their free time [LIT-08]. A comparable UK survey of over 70,000 young people found reading enjoyment sitting at a 15-year low, with the gap between disadvantaged and other children widening [LIT-10]. We should also be honest about where the evidence for fiction reading's cognitive benefits gets shakier: claims that fiction reading builds theory-of-mind or empathy rest on contested ground, and the most methodologically careful test — a genuine two-year longitudinal study of 236 UK adolescents — found fiction reading was associated with theory-of-mind scores at the same time point, but reading at 11 did not predict theory-of-mind at 13. No durable effect, just a same-time association that could run in either direction [LIT-11].
Put together, this evidence doesn't support treating reading as a soft, enrichment-shelf activity that happens after the 'real' academic work is done. A third of Australian children can't read proficiently, and that gap tracks Indigenous status, remoteness and socioeconomic background far more than it tracks ability [RDG-01][RDG-15][RDG-17]. Decoding instruction — done early, done systematically — has a real, moderate, well-replicated effect in the United States and United Kingdom research above, even if the exact number carries an honest asterisk [RDG-06][RDG-07]. But decoding is not the whole job: comprehension depends on background knowledge and vocabulary that themselves depend on how much a child reads [RDG-08][VOC-10][RDG-18], and reading volume is savagely unequal from an early age, compounding across years into differences measured not in percentage points but in multiples — again on United States data [RDG-09]. That compounding is exactly why reading habit-building deserves to be treated as literacy infrastructure rather than enrichment — it is the mechanism through which small early advantages or disadvantages become large, durable ones, decades before anyone measures the endpoint.
Figure · from the registerReading equity, NAPLAN 2024
Two of the sharpest gaps in Australian reading data: the share of students needing additional support (Indigenous vs non-Indigenous) and the share reaching Strong or above (very remote vs major-city schools). Redrawn from #RDG-15.
Figure · from the registerOut-of-school reading volume — the Matthew effect
Words read per year outside school by the average child at the 90th vs 10th percentile of reading volume — a more-than-200-fold gap. These are extrapolated US estimates, not measurements of Australian children. Bar widths are on a square-root scale so the smaller bar remains visible at all; the true ratio is stated. Figures from #RDG-09.
A meta-analysis of the National Reading Panel's phonics database found systematic phonics instruction produces a moderate overall effect on reading (approximately d = 0.4-0.44 across 66 treatment-control comparisons from 38 experiments), with substantially larger effects when phonics instruction begins early (Kindergarten/Year 1, d = 0.55) than when it begins after Year 1 (d = 0.27).
There were 38 studies from which 66 treatment-control group comparisons were derived. ... The mean overall effect size produced by phonics instruction was moderate in size and statistically greater than zero, d = 0.44. ... [Effect sizes for beginning phonics instruction in kindergarten or 1st grade were] significantly larger (d = 0.55) than effect sizes above 1st grade (d = 0.27). These results support Chall's (1967) assertion that early instruction in systematic phonics is especially beneficial to growth in reading.
In the classic 'baseball study', prior topic knowledge had a stronger effect than general reading ability on Year 7-8 students' comprehension and recall of a text: weaker readers with high baseball knowledge performed as well as strong readers, and there was a significant main effect for prior knowledge on all recall measures with no interaction between prior knowledge and reading ability.
We investigated how prior knowledge influences the amount of short-term nonverbal and verbal memory and long-term retention in students of high and low ability in reading comprehension. Sixty-four junior high students were divided into four equal-sized groups on the basis of preassessed reading ability (high and low) and preassessed amount of existing prior knowledge about baseball (high and low). ... There was a significant main effect for prior knowledge on all measures. No interactions between prior knowledge and ability were found. These results delineate the powerful effect of prior knowledge.
Cunningham and Stanovich's research found enormous differences in independent reading volume translate into vocabulary/word-exposure gaps: the average child at the 90th percentile of reading volume reads almost 2 million words per year outside school — more than 200 times the ~8,000 words per year read by the child at the 10th percentile — meaning a full year of a low-volume reader's out-of-school reading equals about two days of a high-volume reader's. This is a US research synthesis of extrapolated estimates from reading-rate and time-reading data, not Australian data and not a single controlled trial.
The average child at the 90th percentile reads almost two million words per year outside of school, more than 200 times more words than the child at the 10th percentile, who reads just 8,000 words outside of school during a year. To put it another way, the entire year's out-of-school reading for the child at the 10th percentile amounts to just two days reading for the child at the 90th percentile!
Nagy and Anderson's classic corpus analysis estimated there are approximately 88,500 distinct word families in printed school English, and argued that direct vocabulary instruction alone could not plausibly account for the number of words children actually learn — implying most vocabulary growth must come through wide reading and context/morphology.
Concludes that there are about 88,500 words in printed school English and that even systematic direct vocabulary instruction could not account for a significant proportion of all the words children actually learn, nor cover more than a modest proportion of the words they will encounter in school reading materials.
A definitive randomised trial of reciprocal reading -- the best known structured comprehension-strategy programme -- in 20 English high schools with 782 Year 7 students found no significant effect on any outcome measure, after two smaller trials of the same programme had reported gains of +0.13 to +0.25. Effects across the eight primary and secondary measures ranged from -0.09 to +0.13, none of them significant, and a sensitivity analysis restricted to students who received the full recommended dose found no effect either.
The trial was conducted with students in 20 English high schools. Process data found the technique implementable by schools. Test data (n=733) found no significant differences between treatment and control groups, although the observed changes in treatment group students was of similar magnitude to those detected in a previous study. ... In total 782 students from 20 recruited schools were randomised, 388 students to the treatment group and 394 students to the control group ... ANCOVA indicated no significant differences in gains between control and intervention groups (F(1730)=0.12, p=0.73. The effect size intervention to control was -0.05 (95 % CI -0.05 to -0.05). ... There were no significant differences between control and intervention groups on any of the primary analyses. It is therefore possible to conclude that being on-treatment did not result in significant gains (or losses) for the Reciprocal Reading treatment group when compared to the control group.
Grattan Institute estimates that roughly a third of Australian children cannot read proficiently, equating to about 8 students in a typical classroom of 24.
Australia has a reading problem. A third of our children can't read proficiently. ... In the typical Australian school classroom of 24 students, eight can't read well.
Grattan Institute's central policy recommendation is a national commitment (the 'Reading Guarantee') to lift the proportion of Australian students who become proficient readers to at least 90%, via a nationally consistent structured-literacy approach including a Year 1 Phonics Screening Check.
All state and territory governments, and Catholic and independent school sector leaders, should commit to a six-step 'Reading Guarantee', pledging that at least 90 per cent of Australian students will become proficient readers.
The US National Reading Panel (2000) concluded that systematic and explicit phonemic awareness instruction directly causes improvements in children's reading and spelling, and recommended it as a core classroom component for pre-school children and beyond who have not yet mastered phoneme concepts.
Many difficulties learning to read were caused by inadequate phonemic awareness ... systematic and explicit instruction in phonemic awareness [is] an important component of classroom reading instruction for children in preschool and beyond who have not been taught phoneme concepts or who have difficulties understanding that the words in oral language are composed of smaller speech sounds.
The US National Reading Panel (2000) found phonics instruction produces significant reading benefits for children from kindergarten through Year 6, and for children with reading difficulties, with systematic phonics instruction producing the strongest effects. This is a US evidence base covering kindergarten to Year 6, not Australian data and not secondary or adolescent readers.
The research literature provides solid evidence that phonics instruction produces significant benefits for children from kindergarten through 6th grade and for children having difficulty learning to read ... systematic phonics instruction is appropriate for routine classroom instruction.
The UK Education Endowment Foundation's Teaching & Learning Toolkit rates phonics approaches as having a positive impact on early reading, worth an average of five additional months' progress, based on extensive evidence, and finds the approach particularly beneficial for pupils from disadvantaged backgrounds.
Phonics approaches are consistently effective in supporting younger pupils to master the basics of reading, with an average impact of an additional five months' progress. ... Phonics has a positive impact on average (+5 months) with extensive evidence and is an important component in the development of early reading skills particularly for children from disadvantaged backgrounds.
In PIRLS 2021, Australia's mean Year 4 reading score (540) was statistically similar to 2016 (544) and higher than 2011 (527); about 80% of Australian students met the PIRLS intermediate/proficient benchmark, with Australia outperformed by 6 countries, similar to 8, and ahead of 28.
Australia's average score of 540 points in PIRLS 2021 was statistically similar to the average score in PIRLS 2016 (544 points) and remains higher than the average score for 2011 (527 points). ... 80% of Australia's students achieved the PIRLS intermediate benchmark. ... 6 countries performed at a higher level than Australia, 8 countries performed at a similar level to Australia, and 28 countries performed at a lower level than Australia.
In PIRLS 2021, Australian girls outperformed boys in reading: 84% of girls met or exceeded the intermediate benchmark compared with 77% of boys, a 7-point gap.
77% of Australian boys and 84% of Australian girls met or exceeded the PIRLS Intermediate benchmark.
Australian 15-year-olds' PISA scores in mathematics AND reading declined by more than 25 points between the early 2000s and 2022 — the source states the figure for the two subjects jointly, not for reading in isolation — a fall equivalent to roughly one year of schooling, meaning 2022's 15-year-olds scored at the level 14-year-olds did two decades earlier, though performance was stable between 2015/2018 and 2022.
While the most recent trend is stable, at least since 2015, over the longer term performance declined in all three subjects. In mathematics and reading, students in Australia scored more than 25 points below students of the same age who sat the PISA test in the early 2000s: in other words, 15-year-olds in 2022 scored at a level that would have been expected of 14-year-olds, 20 years earlier.
In PISA 2022, 79% of Australian 15-year-olds attained at least Level 2 (baseline) reading proficiency (above the 74% OECD average), and 12% were top performers at Level 5 or higher (above the 7% OECD average); compared with 2012, the share of Australian students below Level 2 in reading rose by 6 percentage points.
Some 79% of students in Australia attained Level 2 or higher in reading (OECD average: 74%). ... In Australia, 12% of students scored at Level 5 or higher in reading (OECD average: 7%). ... Compared to 2012 the proportion of students scoring below a baseline level of proficiency (Level 2) increased by six percentage points in mathematics; by six percentage points in reading; and by five percentage points in science.
NAPLAN 2024 national results show, in reading, an average of 67.0% of Australian students nationally achieved 'Exceeding' or 'Strong' proficiency (ranging from 66.3% in Year 3 up to 71.4% in Year 5, then down to 63.0% in Year 9), while 10.3% were in the 'Needs additional support' category on average.
In reading, the average proportion of students who achieved at "Exceeding" and "Strong" levels in 2024 was 67.0%, increasing from Year 3 (66.3%) to Year 5 (71.4%), then dropping in Year 7 (67.3%) and in Year 9 (63.0%). In reading, the average proportion of students who achieved at "Needs additional support" levels in 2024 was 10.3%, decreasing from Year 3 (11.3%) to Year 5 (8.7%), and increasing in Year 7 (10.2%) and Year 9 (11.1%).
NAPLAN 2024 shows a large equity gap in reading: nationally, around 1 in 3 Indigenous students (33.7%) were in the 'Needs additional support' reading category compared with fewer than 1 in 10 non-Indigenous students (8.6%); students in very remote schools were far less likely to reach 'Strong'/'Exceeding' reading levels (24.0%) than students in major cities (70.7%).
In reading, around one in 3 of Indigenous students (33.7%) are in the "Needs additional support" level while fewer than one in 10 of non-indigenous students (8.6%) fall into that proficiency level. ... in reading, 24.0% of students from very remote schools are in this [Strong/Exceeding] category, compared to 70.7% of students from major city schools.
OECD PISA 2009 analysis found reading for enjoyment daily is associated with substantially better reading performance: on average, students who read daily for enjoyment scored the equivalent of about one-and-a-half years of schooling better than those who did not, though the proportion of students reading daily for enjoyment fell in most OECD countries between 2000 and 2009 (from 69% to 64% on average).
PISA finds that a crucial difference between students who perform well in the PISA reading assessment and those who perform poorly lies in whether they read daily for enjoyment, rather than in how much time they spend reading. On average, students who read daily for enjoyment score the equivalent of one-and-a-half years of schooling better than those who do not. ... On average across OECD countries, the percentage of students who reported reading daily for enjoyment decreased by five percentage points during the period. In 2000, 69% of students reported reading for enjoyment daily, but in 2009, only 64% of students did so.
OECD PISA 2009 data show reading-for-enjoyment is unevenly distributed: socio-economically advantaged 15-year-olds were much more likely to read daily for enjoyment (72% on average across OECD) than disadvantaged students (56%), with Australia among ten OECD countries where this socio-economic gap exceeded 20 percentage points.
On average across OECD countries, 72% of socio-economically advantaged students ... reported that they read daily for enjoyment while only 56% of disadvantaged students reported doing the same. ... In ten OECD countries — Australia, Austria, Belgium, Estonia, France, Germany, Ireland, Korea, Luxembourg and Switzerland — the difference in the share of socio-economically advantaged and disadvantaged students who read for enjoyment is more than 20 percentage points.
In PISA 2018, 36% of Australian 15-year-old girls and 53% of Australian 15-year-old boys reported they do not read for enjoyment — up 11 and 13 percentage points respectively from PISA 2000 (25% and 40%). Agreement with 'I read only if I have to' rose from 30% to 46% among Australian girls and from 46% to 61% among boys over the same 2000-2018 span.
Between 2000 and 2018, the proportion of Australian girls who reported they did not read for enjoyment increased by 11 percentage points, to 36 per cent. The proportion of Australian boys who did not read for enjoyment increased by 13 percentage points, to 53 per cent. ... The proportion of girls who reported 'I read only if I have to' increased from 30 per cent to 46 per cent, and from 46 to 61 per cent for boys.
In PISA 2025, 77% of Australian 15-year-olds attained at least Level 2 (baseline) reading proficiency (OECD average 69%), and 11% scored at Level 5 or higher (OECD average 6%); both Australian shares are lower than the 79% and 12% reported for PISA 2022, while the OECD averages fell further (74% to 69%, 7% to 6%), so Australia's margin over the average widened. Reading was a minor domain in the 2025 cycle and the note prints no significance test for the change in shares.
Some 77% of students in Australia attained Level 2 or higher in reading (OECD average: 69%). ... In Australia, 11% of students scored at Level 5 or higher in reading (OECD average: 6%).
ACARA's NAPLAN Assessment Framework sets a target range of 0-15% of the Years 3 and 5 reading test for the cognitive process 'analysing and evaluating', rising to 5-30% at Years 7 and 9 -- so the reading test is weighted overwhelmingly toward locating, identifying, integrating and interpreting, and a test may contain no analysis items at all in the younger years.
Table 8: NAPLAN reading Years 3 and 5 test content ... Cognitive process Target range / Locating and identifying 30–60% / Integrating and interpreting 35–60% / Analysing and evaluating 0–15% ... Table 9: NAPLAN reading Years 7 and 9 test content ... Cognitive process Target range / Locating and identifying 15–45% / Integrating and interpreting 40–65% / Analysing and evaluating 5–30%
An Australian-led meta-analysis of 106 school-based reading interventions covering 156,981 students in Grades 4-12 found an aggregate effect of g = 0.212 on reading outcomes, with the strongest effects for vocabulary, and reported that briefer interventions of 15 hours or less and interventions measured with researcher-developed rather than standardised tests tended to show stronger effects.
Peer-reviewed articles and reports published between 2011 and 2023 (in English) were eligible if they described a school-based intervention addressing and assessing one or more reading skill(s), targeted Grade 4-12 students with or at risk of reading difficulties, and included a control or comparison group (k = 106; n = 156,981). Correlated and Hierarchical Effects meta-analyses with robust variance estimation yielded an aggregate effect of g = 0.212 on reading outcomes, with a Moderate to High level in quality of evidence. Subgroup and moderator analyses indicated that the impact of intervention varied by reading outcome, with the strongest effects shown for vocabulary. ... Results published more recently and in academic journals tended to show stronger effects overall, as did briefer interventions ('15 h), and interventions assessed using bespoke researcher-developed measures.
The UK National Literacy Trust's 2017/18 Annual Literacy Survey of 49,049 children and young people aged 8 to 18 found that nearly twice as many children aged 8 to 11 as those aged 14 to 16 said they enjoy reading, and twice as many said they read daily in their free time.
49,049 children and young people aged eight to 18 in the UK participated in our survey between November 2017 and January 2018. ... There are also differences by age, with nearly twice as many children aged 8 to 11 as those aged 14 to 16 saying that they enjoy reading ... twice as many children aged 8 to 11 as those aged 14 to 16 saying that they read daily in their free time.
Cite any finding by its anchor — for example #RDG-06 — and the link resolves to the claim, quote and source above. The verification protocol lives on the register’s main page.
