Comprehension strategies, and what outranks them.

The finding most lists omit

In a 1988 study that has been replicated in spirit many times since, researchers gave Year 7–8 students a passage about a baseball game. Students were grouped by reading ability and by how much they knew about baseball. Prior topic knowledge had a stronger effect than general reading ability on comprehension and recall: weaker readers who knew baseball performed as well as strong readers who did not RDG-08.

This is the single most useful thing a parent can know about comprehension, because it explains the pattern that otherwise looks like inconsistency. A child who follows one text easily and stalls on the next is usually not reading badly on the second. The second text assumes something the first did not.

Where the knowledge comes from

Printed school English contains roughly 88,500 distinct word families— an American corpus analysis, and the closest estimate anyone has published — and it argued that even systematic direct instruction could not account for a significant proportion of the words children actually learn, nor cover more than a modest proportion of the words they will meet RDG-18. Something else does most of the work, and the volume figures show what.

More than two hundred times the exposure. Put the other way round, a full year of the low-volume reader’s out-of-school reading is about two days of the high-volume reader’s. These are extrapolated estimates from a United States research synthesis rather than measurements of Australian children, and they are quoted here for the scale of the gap rather than the precision of either number RDG-09. An OECD analysis across its member countries found that reading daily for enjoyment was associated with reading performance about one and a half years of schooling ahead of students who did not — an association in survey data, not a demonstration that reading causes the gain — and that what mattered was reading daily for enjoyment rather than the number of hours spent RDG-16.

The strategies, kept short on purpose

These are worth teaching and worth practising. They are listed briefly because the argument above is that they are the smaller half, and a page that padded them to listicle length would be contradicting itself.

1. Read for the question the text is answering. Before reading closely, work out what problem or question the passage exists to address. A text held together by its question is far easier to hold in mind than one held together by its order.
2. Say the paragraph back in one sentence. After each paragraph, state its job in a sentence — not a summary of its content, but what it does for the argument. Where that sentence will not come, comprehension has already broken and rereading is worth it.
3. Predict, then check. Guess where the text is going, then read on to see. Being wrong is the useful case: a failed prediction marks precisely where your model of the text and the text itself part company.
4. Notice the words doing the joining. However, therefore, although, in spite of. These carry the logical shape of a passage, and readers who skim them lose the argument while still following the topic.
5. Build the knowledge the text assumes. When a passage is hard because of what it takes for granted rather than how it is written, the fix is the subject, not the technique. This is the strategy the research puts first, and the one least often taught as one.

There is a reason to be modest about that list beyond the argument above, and it is worth stating because the strategies-first approach is usually presented as settled. Reciprocal reading is the best known structured programme built on these strategies. It was trialled properly in 20 English high schools, with 782 Year 7 students randomised individually, after two smaller trials had reported gains of between +0.13 and +0.25. The definitive trial found nothing — on eight separate measures across two different test batteries, with effects running from −0.09 to +0.13 and not one of them significant — and nothing again when the analysis was narrowed to the students who had received the full recommended dose. These were English schools rather than Australian ones, and the trial ran through COVID-19 disruption, which its authors say may have produced atypical findings RDG-24.

Take that for what it is. It is one programme, and it is not evidence that teaching comprehension is pointless. What it does show is how a routine earns an evidence-based label from early trials and then fails the decisive one — and why this page spends more words on what a reader knows than on what a reader is told to do.

How common the difficulty is

The Grattan Institute estimates about a third of Australian children cannot read proficiently — roughly eight students in a classroom of 24 RDG-01. Over the longer run, Australian 15-year-olds’ PISA scores fell by more than 25 points between the early 2000s and 2022 — about a year of schooling — though performance was stable between 2015 and 2022. The OECD states that figure for mathematics and reading together rather than for reading alone, so it is a measure of how far the cohort has moved overall, not a reading-specific decline RDG-12.

The most recent cycle puts that in better proportion. In PISA 2025, 77% of Australian 15-year-olds reached at least Level 2 in reading against an OECD average of 69%, and 11% reached Level 5 or higher against an OECD average of 6% RDG-21. Both Australian shares are slightly below the 79% and 12% reported for 2022, but the OECD averages fell further, so the margin over the average widened. Read that carefully rather than either way it could be spun: reading was a minor domain in the 2025 cycle, the country note prints no significance test for the change, and a one-cycle movement of two points is not a trend.

These figures are worth holding lightly in the individual case. They describe a national distribution, not a particular child, and a child who reads slowly is not thereby one of the eight.

Where these claims come from

Every figure above is drawn from the reading domain of the Evidence Register, where each claim appears with its verbatim quote and full citation. The vocabulary domain carries the word-knowledge research this page touches on.

Working on it with someone

Argo Academics is a one-to-one English tutoring practice in Melbourne, where Callum Hay teaches every lesson personally. Reading closely is most of what the lessons do — and where a text is hard because of what it assumes rather than how it is written, that is visible in a way it never is on a worksheet.

If you are not sure whether that is the problem, the two-minute diagnostic asks what you have actually seen at home rather than what a score says — which is the more useful evidence at this stage.