Guides · Reading
Every Reading question type on the NSW Selective Test, with examples.
Reading is one of the four sections of the NSW Selective Test, worth 25% of the total score. In its current form it is 45 minutes and 38 answers, spread across 17 on-screen questions — three of which are multi-part — organised into five distinct parts. It is sat entirely on screen. Each part tests a different way of reading, and the last three parts don't work like ordinary multiple choice at all. Below is what every part asks for, with one fully worked example each.
This describes the current format
For the May 2026 test the Reading section was rebuilt: it moved from 30 questions in 40 minutes to the five-part, 38-answer format described here. A lot of the practice material still circulating online describes the retired shape. Everything on this page is the format your child will actually sit.
The five parts at a glance
Unlike Mathematical Reasoning or Thinking Skills, the Reading section isn't a flat run of same-shaped questions — it's five fixed parts, each with its own passage type and its own answering mechanic. The question counts are stable from paper to paper.
| Part | Answers | What it tests | How you answer |
|---|---|---|---|
| 1 · Fiction | 8 | Comprehension of one prose passage — detail, inference, tone, technique. | One question at a time, four options. |
| 2 · Vocabulary cloze | 8 | The word that fits a gap, decided by collocation and precise meaning. | One grouped screen of gaps. |
| 3 · Poetry | 6 | Imagery, figurative language and implied meaning in a complete poem. | One question at a time, four options. |
| 4 · Gapped text | 6 | How a text holds together — slotting removed extracts back in by cohesion. | Shared bank of extracts, one decoy. |
| 5 · Multiple matching | 10 | Matching each statement to the right text among several short texts. | A small set of labels, reused across all ten. |
8 + 8 + 6 + 6 + 10 = 38 answers. Parts 2, 4 and 5 each appear as a single multi-part screen, which is why the on-screen count is 17 while the answer count is 38.
The most important thing to notice is that more than half of the answers — Parts 2, 4 and 5, twenty-four of the thirty-eight — don't give each question its own fresh set of options. Instead a small pool of choices is shared across many gaps or statements. A child who has only practised classic one-question-one-answer multiple choice can lose time simply working out how these parts are meant to be answered. That is the single biggest reason to see the real format before test day.
Part 1 · Fiction comprehension (8 questions)
A single prose extract — usually a scene from a story — followed by eight questions, each with four options. This is the most familiar part, but "familiar" doesn't mean easy: the questions climb from literal detail up to inference the passage only implies. The types within it:
- Literal detail. Something the passage states directly — the answer is there, but the wrong options quote real phrases from elsewhere in the text to tempt a hasty reader.
- Character inference. What a character feels or wants, worked out from what they do and say rather than any label the author gives.
- Implied meaning. A conclusion the passage points to without stating — the highest-value type, and the one most practice material handles badly.
- Tone and attitude. The narrator's or a character's stance — wry, uneasy, admiring — read from word choice.
- Technique and effect. Why the author made a particular choice — a repeated image, an abrupt sentence — and what it does to the reader.
- Text structure. How the passage is organised — a shift in time, a turn in the argument, the function of a particular paragraph.
Worked example — implied meaning
Nour had practised the piece for six weeks, and now, seated at the piano with the hall full, she found her hands perfectly still. She looked at the keys as if they belonged to someone else. When Mr Achterberg nodded from the wings, she did not begin. She counted the exit signs instead — one, two, three — and thought about how quiet the car park would be.
What does the passage most strongly suggest about Nour?
Answer: C.
Why: the passage never says Nour is frightened — but every detail points there, and the decisive ones are that she counts the exit signs and pictures the empty car park. She is rehearsing her escape, not her performance. Option B mistakes her stillness for composure (the passage frames it as frozen, not relaxed); A takes "did not begin" too literally as forgetting; D invents an emotion the text never supports. Notice the pattern: each wrong option is half-supported by one detail read on its own — which is exactly how this part is built to catch a reader who stops at the first plausible answer.
Part 2 · Vocabulary cloze (8 gaps)
A short passage with eight words removed. For each gap the child chooses the word that fits — and the options are always near-synonyms, so a dictionary definition won't separate them. What decides the answer is collocation (which words naturally keep company) and precise shade of meaning. This is not a spelling or vocabulary-list test; it rewards a child who has read widely enough to have an ear for which word simply sounds right.
- Collocation. All four options mean roughly the same thing, but only one is the word English habitually pairs with the surrounding phrase.
- Word in context. The gap turns on a fine distinction in meaning that only becomes visible from the sentence around it.
Worked example — collocation (one gap shown)
Despite the storm, the old bridge remained ______, carrying traffic as it had for a century.
Answer: A.
Why: all four words carry the idea of "not broken, all there," so a definition can't split them. But only intact collocates with a structure surviving damage — a bridge "remained intact" through a storm. "Entire," "whole" and "complete" describe totality (a whole cake, the complete set), not the sense of having come through undamaged. The gap is decided by fit, not by meaning in isolation — which is the whole point of this part.
Part 3 · Poetry (6 questions)
A complete short poem, then six questions with four options each. Poetry unsettles a lot of strong prose readers, because the answer almost never sits on the surface — it lives in what an image implies. The types:
- Comprehension. What is literally happening in the poem — who, where, and what changes — read past unusual word order and line breaks.
- Figurative language. What a metaphor, simile or piece of personification is doing — and, crucially, what it suggests, not just what it names.
- Form and effect. How a choice of rhythm, repetition or shape shapes the reader's response.
Worked example — figurative language
The poem describes a river in flood, ending on the couplet:
The river carried the whole town's secrets down,
and never once looked back to see them drown.
The phrase "never once looked back" most likely suggests that the river is
Answer: B.
Why: the line gives the river a human capacity — "looked back" — only to deny it, and that denial is what paints the river as indifferent, carrying things off without a backward glance. Option A reads "carried … down" too literally as speed; C invents a fear the calm tone of the couplet contradicts; D answers a question about the river's shape, not the human quality the line assigns it. Poetry questions turn on what an image implies about a subject, not on what it literally describes.
Part 4 · Gapped text (6 gaps)
An informational passage with six sentences or short extracts removed, and a shared bank of extracts to slot back into the right gaps. The bank always holds one more extract than there are gaps, so at least one option is a decoy that fits nowhere. This part isn't testing whether a sentence is true or on-topic — it's testing cohesion: does the extract you choose link what comes before the gap to what comes after it?
- Gapped-text insertion. Choose the extract whose meaning and connective words (this, however, as a result) bridge the sentences on either side of the gap.
- Paragraph-summary matching. The same cohesion skill applied at paragraph scale — matching a one-line summary to the paragraph it captures.
Worked example — gapped-text insertion (one gap shown)
Lighthouses once needed a keeper who lived on site all year round. ______ Today, sensors detect a failing bulb and switch to a backup automatically, and a technician visits only twice a year.
Which extract best fills the gap?
Answer: B.
Why: the gap sits between a "before" (a keeper lived on site) and an "after" (today, sensors do it automatically). Only B carries the paragraph across that divide — it names the turning point, automation, that explains the shift. Options A and C are both perfectly true and on-topic, but each just adds a detail; neither links the two states the paragraph is moving between. A gapped-text option is judged on the join it makes, not on whether it's a reasonable sentence about lighthouses — and the decoy in the bank is always a reasonable sentence that makes no join.
Part 5 · Multiple matching (10 questions)
Several short texts, each labelled — say, four accounts labelled A to D — and ten statements. For each statement the child finds the one text it best fits, and the same labels are reused across all ten statements. It's the longest part, and it rewards a particular discipline: matching on the statement's precise condition, not merely its topic.
- Statement-to-text matching. Locate the text that satisfies exactly what the statement claims — a change of mind, a specific reason, a comparison — when several texts touch the same subject.
- Cross-text discrimination. Hold four short texts in mind at once and tell apart near-neighbours that agree on the topic but differ on the detail the statement actually asks about.
Worked example — statement-to-text matching
Four students each describe a school excursion (texts A–D). Which student changed their opinion during the day?
Answer: B.
Why: the statement's condition is a change of opinion. Only B moves from one position ("sure I'd hate it") to its opposite ("joined the club"). A and D each hold a negative view steady; C holds a positive one steady — no change in any of them. The trap this part sets is a text that shares the statement's topic (all four are about liking or disliking the trip) while failing its precise condition. Because A–D are reused across every statement, you're always choosing the best-fitting text, never picking from a fresh list.
Where this breakdown comes from
This is our own analysis. We classified 114 questions across four official Reading papers — the Department's three practice tests for 2026 entry and the March 2022 live paper — question by question, into the five parts and the reading skills within them. The five worked examples above are original, written by us to illustrate each part's mechanic — not copied or reworded from any published question. Our Reading practice papers are being built to this same map, in the same part structure and proportions the official papers use.
The format facts — 45 minutes, 38 answers, five parts, sat on screen — are current for the May 2026 test onward and match the section format published by the Department. The wider test shape is set out in the NSW Selective Test guide.