11 Plus Comprehension: Question Types and How to Answer Them

Most guides to 11 plus comprehension name the same four or five question types and then stop. Your child learns that retrieval questions exist, that inference is important, that vocabulary matters and that authors use techniques. What almost nobody shows is the thing that actually decides the score: what the person holding the pen is looking for, and how many separate things they need to see before they can award full marks.

That gap matters, because a large share of lost comprehension marks are not comprehension failures at all. The child understood the passage perfectly well. They wrote one thoughtful sentence for a three-mark question, or paraphrased when the question said copy, or ticked three boxes when the instruction allowed two. Those are format errors, and format errors are much faster to fix than reading ability is to build.

So this guide is organised as a series of decisions rather than a list of labels. Decision one: which question type am I looking at, judged from the wording alone. Decision two: how many separate points does this answer need, judged from the mark tariff. Decision three: on this particular paper, do I read first or look at the questions first. Get those three right and the marks follow, because the reading was rarely the problem.

What 11 plus comprehension actually tests, and what it does not

Grammar school and independent school papers vary a great deal by region, but nearly all of them are built on the same national foundation: the Key Stage 2 reading assessment framework. That document is the closest thing to an official specification for what an eleven-year-old is expected to do with a text, and it is public.

The national test itself is a single component worth 50 marks in 60 minutes including reading time, with a reading booklet and a separate answer booklet, and a selection of texts totalling roughly 1,500 to 2,300 words across fiction, non-fiction and poetry (KS2 English reading test framework). Texts are ordered by increasing reading demand, and questions are set in order of difficulty where possible while maintaining chronology with the text. That last detail is quietly useful: if your child is on question 7 and hunting near the start of the passage, they are probably in the wrong place.

Just as important is what comprehension does not test. It is not a memory test and it is not a general knowledge test. The Essex CSSE paper states this outright to candidates: "REMEMBER, this is not a test of memory. You can look back at the passage to check your answers as many times as you want" (CSSE familiarisation English paper). Children who answer from a general impression of the passage, rather than going back and physically locating the words, are throwing marks away by choice.

Where the marks really are

Here is the first thing competitors will not tell you, because it requires reading a specification rather than writing tips. The framework publishes the share of marks that each content domain can carry, and the spread is enormous.

• 2b, retrieve and record information or identify key details: 8 to 25 marks (16 to 50%)
• 2d, make inferences and justify them with evidence: 8 to 25 marks (16 to 50%)
• 2a, give or explain the meaning of words in context: 5 to 10 marks (10 to 20%)
• 2c, summarise main ideas from more than one paragraph: 1 to 6 marks (2 to 12%)
• 2e, predict what might happen: 0 to 3 marks (0 to 6%)
• 2f, how content contributes to meaning as a whole: 0 to 3 marks (0 to 6%)
• 2g, how meaning is enhanced through word choice: 0 to 3 marks (0 to 6%)
• 2h, make comparisons within the text: 0 to 3 marks (0 to 6%)

Retrieval and inference can between them account for half or more of the paper. Prediction can account for nothing at all. If your child's revision time is being split evenly across eight skills, it is mis-allocated by a wide margin.

The tariff distribution reinforces the point. Of the 50 marks, 22 to 33 sit on one-mark questions, 10 to 20 on two-mark questions and only 3 to 12 on three-mark questions. Format weightings follow: short constructed response (find and copy, short answers) carries 40 to 60% of the paper, selected response 10 to 30% and extended response 20 to 40%.

The revision priority order that falls out of this is unambiguous. Drill accurate retrieval until it is automatic. Then drill inference with evidence. Then vocabulary in context. Treat prediction, comparison and whole-text structure as polish, not as a syllabus.

Decision one: naming the question type from its wording

A child who can classify a question in three seconds has already chosen the answer shape. Train the trigger words, not the theory.

• "Find and copy", "Give one word", "According to the text", "Name two things" signals retrieval. Answer shape: lift the exact words.
• "Why do you think", "How can you tell", "What does this suggest", "Explain, using evidence from the text" signals inference. Answer shape: point, evidence, explanation.
• "What does the word X mean in this line", "Which word closest in meaning to" signals vocabulary in context. Answer shape: a substitute word that fits the sentence.
• "How does the writer", "Why has the author chosen", "What is the effect of" signals technique. Answer shape: several separate labelled features.
• "Summarise", "What are the main points", "How is X different from Y" signals summary or comparison. Answer shape: coverage across paragraphs.

The mark number in the margin is the second half of the signal. CSSE tells candidates plainly that "the number of marks available for each question is indicated in the right hand margin", and that where they must choose between possible responses they should "choose always the MOST APPROPRIATE response".

Retrieval questions: find and copy, never paraphrase

Retrieval is the largest single pot of marks and the easiest to protect. The rule is brutal and simple: if the question says find and copy, copy. Do not improve, shorten or tidy the author's words. A child who writes "he was scared" when the text says "his hands would not keep still" has demonstrated understanding and lost the mark, because the question asked for retrieval and got inference.

Two further habits protect these marks. First, answer in the number of items requested: "name two" wants two, and a third guess can invalidate the response. Second, check spelling when copying. The CSSE mark scheme is explicit on one vocabulary item that "SPELLINGS MUST BE CORRECT" (CSSE marking sheet), and a word copied from a passage sitting open in front of you is not a spelling many markers will forgive.

11 plus inference questions: point, evidence, explanation

Inference carries as much of the paper as retrieval, and it is where technique separates children most sharply. The framework quietly tells you why: it lists "make developed inferences drawing on evidence from the text" and "explain and justify inferences, providing evidence from the text to support reasoning" as two separate expectations. The inference and the justification are distinct demands, which is exactly how 11 plus inference questions tend to be marked.

So the answer shape has three moving parts:

1. Point: state the inference in a plain sentence. The narrator does not trust the visitor.
2. Evidence: quote or reference the words that prove it. She "kept the chain across the door" while they spoke.
3. Explanation: say what the evidence shows. Leaving a barrier in place during a conversation suggests she expects to need protection.

And there is a trap worth naming. A CSSE inference question requires "some reference to miracle or reference to lines 5/6, not just that he had taught a cat to speak": in other words, restating the obvious plot event scores nothing. If the answer to "why is this surprising" is simply a description of what happened, no inference has been made. The child must say what the event implies.

Vocabulary in context: substitute, do not define

Vocabulary questions are worth 10 to 20% of the marks and are lost mostly to dictionary instinct. The child produces a general definition, when the question asked what the word means in this line.

Teach substitution instead. Cover the word, read the sentence, and decide which word would keep the sentence meaning the same. If the text says the crowd "drifted" out of the hall, the answer is "moved slowly and aimlessly", not "floated on water". Then read the substitute back into the original sentence as a check. If it sounds wrong, it is wrong.

Technique, summary and comparison: the higher-tariff group

These are the questions where children who read well still score two out of four. The framework's own example of an extended response is "How does the writer increase the tension throughout this paragraph? Explain fully referring to the text in your answer." The word fully is doing real work there.

The CSSE mark scheme for an author-technique question is marked "Three of:" against a list of possible features, including elaborate vocabulary and long words, long or complicated sentences, formality of tone and the choice of names. Three separate features, three marks. One beautifully developed paragraph about a single feature cannot reach full marks no matter how well written it is.

For summary and comparison, coverage is the same currency. Summary questions ask for main ideas from more than one paragraph, so an answer drawn entirely from paragraph one is structurally incomplete before the marker reads a word of it.

Decision two: how many points does this answer need?

This is the single highest-value habit in the whole subject, and it takes one week to install.

The framework states that "for multi-mark questions, if the correct answer is not reached and, therefore, full marks cannot be awarded, the mark scheme will provide details of how partial credit can be awarded". Marks are built up in identifiable pieces. CSSE's paper shows what those pieces look like: a metaphor question splits one mark for writing out the metaphor and one mark for explaining what it means. Identify, then explain. Two marks, two actions.

So the rule for your child is mechanical: read the number in the margin, then make that many separate moves. Two marks, two moves. Three marks, three moves, and preferably three lines so the marker can see them. The framework also notes that on open two- and three-mark questions markers should consider the pupil's ability to express reasoning through developed answers, so the points do need substance. But a developed answer that contains one idea is still one idea.

A worked passage: underline and locate

Here is a short original passage with numbered lines, in the style of papers that number lines for exactly this purpose.

1. The kitchen at Marran Hall had not been cleaned since the last tenant left,
2. and Nell went at it with a determination that surprised even her mother.
3. She scrubbed the range until her knuckles split, then wrapped them in a cloth
4. and carried on. By four o'clock the room smelled of soda and wet stone.
5. Her mother stood in the doorway with the letter still unopened in her hand.
6. "We may not be staying," she said. Nell did not look up. She began the floor.

Three original questions on it:

• (a) Find and copy the phrase that shows how the kitchen smelled by four o'clock. (1 mark)
• (b) What does Nell's behaviour in lines 5 and 6 suggest about how she is feeling? Use evidence from the text. (2 marks)
• (c) How does the writer show that Nell is determined? Give three ways. (3 marks)

Worked answers at full credit and at partial credit

(a) Retrieval, 1 mark. Full credit: "of soda and wet stone". Zero credit: "it smelled clean and damp". The second answer is accurate and demonstrates understanding, and it scores nothing, because the instruction was to copy.

(b) Inference, 2 marks. Full credit: Nell is upset by what her mother has said and does not want to discuss it. She "did not look up" and immediately "began the floor", which shows she is using work to avoid the conversation. That is a point plus evidence plus explanation.

Partial credit, 1 mark: She keeps working and does not look up. This is a correct observation but it is only the evidence half. Nothing has been inferred from it, so the second mark cannot be awarded. This is by far the most common two-mark answer, and it is why children who "understand the story" still score half.

(c) Technique, 3 marks. Full credit needs three separate ways, one line each:

1. She continues after injuring herself: she "wrapped them in a cloth and carried on".
2. The scrubbing is described as physically extreme, until "her knuckles split".
3. She starts a new task straight away rather than stopping, as she "began the floor".

Partial credit, 1 mark: a single well-written paragraph explaining that scrubbing until her knuckles split shows how hard she is working. Fluent, correct, and capped at one mark, because the question asked for three ways and received one.

Decision three: read first or questions first?

Almost every guide asserts that you should read the text first. That advice is right in some rooms and wrong in others, because it depends on the shape of the paper your child is sitting. Check your target schools before drilling a method, and see our comparison of GL, CEM and ISEB formats if you are not yet sure which applies.

Papers with protected reading time: read first. The CSSE English paper allows 10 minutes of reading time before the test begins, then 60 minutes to complete it, with suggested timings of 30 minutes for comprehension, 10 for applied reasoning and 20 for continuous writing. Reading time that does not consume answering time makes read-first structurally correct: use it to read once for meaning and once with a pencil. Our guide to the CSSE paper covers the rest of that structure.

Short passages with multiple-choice questions, GL-style: scan the questions first. When the extract is short and the questions are mostly selected response, the questions function as a search list. Skim them, then read with targets in mind, marking the passage as you find each one.

Online adaptive tests such as the ISEB Common Pre-Test: neither, exactly. The Common Pre-Test is an online adaptive assessment covering maths, English, verbal and non-verbal reasoning (ISEB). There is no paper to annotate and no flicking back and forth, so the method must change: read in chunks, hold the structure of each paragraph in mind, and use on-screen scrolling deliberately rather than guessing from memory.

Whatever the format, the middle step never changes. Underline the words that answer the question before writing anything, so the answer is built from located evidence rather than recollection.

The instruction errors that cost whole questions

This is the tactical layer that no tips post mentions, and it is worth more marks than another month of reading practice.

The CSSE mark scheme opens with a warning: "Candidates must NOT tick more boxes than they are instructed to. Any who do will lose all the marks for that question." The multi-tick questions on that paper are worth three and four marks each. A child who hedges by ticking a fourth box on a four-mark question can lose the lot, not a fraction of it.

Three habits close this off:

• Count the ticks against the instruction before moving on. If it says two, there must be exactly two.
• Never hedge in writing either. Offering two contradictory answers and hoping one lands is usually treated as an incorrect response.
• Move on and come back. CSSE tells candidates to "work quickly but carefully" and, if stuck, to "go on to the next one and return to it if you have time at the end". A three-mark question is not worth the four one-mark questions it can eat.

Punctuation is also assessed in places, with CSSE instructing that "punctuation should be both clear and exact", so full stops and capitals in written answers are not optional decoration.

Turning this into practice that works

The fastest route to improvement is not more passages. It is fewer passages, marked properly.

Sit one passage a week under timed conditions, then mark it together against the tariff rather than against a sense of whether the answer was good. For every question, ask one question: did the answer contain as many distinct moves as there were marks? Keep a tally of losses by cause, splitting them into paraphrased when copy was needed, evidence without explanation, too few points for the tariff, and instruction breaches. Within a month the pattern will be obvious, and it will almost always be concentrated in one or two of those four.

When you want volume, use full papers rather than worksheets so timing pressure is realistic. Trace Learn's paper tests mark answers against tariff-based schemes and show where partial credit was lost, and you can pair them with our free 11 plus practice papers for extra passages. For the wider subject, including the writing task that usually sits alongside comprehension, start with our 11 plus English guide, and when you turn to the paper on the other side of the timetable, work through every 11 plus maths topic and question type in the same way.

Formats differ by region and by school, so confirm the paper your child will actually sit before committing to a method. But the underlying economy is consistent everywhere: most of the marks sit on retrieval and inference, most questions are worth one mark, and multi-mark questions pay out in pieces. Teach a child to read the tariff and answer in that many moves, and comprehension stops being a matter of talent.