A comprehension exercise works when the passage is unfamiliar enough that it cannot be answered from prior knowledge, and the questions ladder from literal retrieval through inference to evaluation. The commonest fault is questions answerable without reading the passage at all.
We do not write the questions. TurtleDock holds no question bank: this gives you the paper’s structure and the marks arithmetic, and the items are set by the subject teacher against the syllabus actually taught.
Comprehension is the most-set and least-examined exercise in a language classroom. Almost every school sets it weekly; very few check whether the questions actually require the passage.
The test is simple and worth running on any comprehension you are about to set: cover the passage and try to answer the questions from general knowledge. Anything you can answer is not testing comprehension.
Three properties, and they trade off against each other.
A good set moves up the ladder. A weak set sits entirely on the bottom rung, which is why students can score well while reading badly.
| Level | What the student does | Example shape |
|---|---|---|
| Literal | Retrieves something stated directly | What did the narrator take with them? |
| Vocabulary in context | Infers meaning from surrounding text | What does "brittle" mean as used here? |
| Inferential | Combines two stated things to reach an unstated one | Why was the narrator reluctant to go back? |
| Structural | Reads how the passage is built | Why does the writer begin with the ending? |
| Evaluative | Judges, with evidence from the text | Is the narrator reliable? Support your answer. |
Literal questions mark themselves. Inferential and evaluative ones do not, and an ambiguous scheme is where comprehension marking becomes inconsistent between teachers marking the same class.
Write the scheme in terms of what the answer must contain — the two pieces of evidence, the link between them — rather than in terms of a model answer. A model answer invites marking by similarity, which penalises students who reached the same understanding by a different route.
Phrased the way they arrive, answered so each one stands on its own.
Short enough that reading it twice fits comfortably in the time allowed. Students who read it once because there was no time are being tested on reading speed rather than comprehension.
Not for assessment. A studied passage tests recall of class discussion. Use unseen passages for testing and textbook ones for teaching.
The passage is the easy part. The questions are the exercise, and laddering them properly against a specific passage for a specific class is exactly the judgement that cannot be outsourced — a generated set sits almost entirely on the literal rung.
Checked by the Assessment desk on 8 September 2026. Corrections go in the page, dated, rather than quietly.
Pricing is on the page, in the open, with the counting rule written next to it. Nobody has to book a call to find out what it costs.
See pricingForty minutes with an implementation engineer — your classes, your fee heads, your grading scheme, entered as we go. No slides, no sample school.