In short
- Six question types cover most of what a reading quiz can ask: literal, reorganisation, inference, prediction, evaluation and personal response.
- Literal and reorganisation questions suit A1 and A2 readers best; inference and evaluation need more support and a simpler form.
- A good quiz mixes types, keeps the text in front of the reader and stays short.
- Type and form are separate choices: any type can be asked as multiple choice, true or false or short answer.
A reading quiz can ask for a fact, a link between two facts, a reason the writer never states, a guess about what happens next, a judgement, or a personal reaction. Each is a different question type, and each checks a different kind of understanding. If every question asks for a fact, you learn only whether pupils can find facts. This article names the types, defines each in one sentence, and shows all six on one short story you can use in class.
What are the main types of reading comprehension question?
The clearest list for language teachers comes from Day and Park (2005), who describe six types of comprehension: literal, reorganisation, inference, prediction, evaluation and personal response. A literal question asks for something stated in the text. A reorganisation question asks the reader to combine facts from different parts of the text. An inference question asks for something the text implies but never states. A prediction question asks what might happen next or after the story ends. An evaluation question asks for a judgement about the text. A personal response question asks how the reader feels about it.
Day and Park (2005) built their list partly on the earlier work of Nuttall, whose handbook, Nuttall (2005), gives a whole chapter to questioning. The six types are a checklist, not a theory of reading. Alderson (2000) warns that it is far from clear whether a given test question measures one separate skill. So use the types to plan a balanced quiz, not to diagnose six separate abilities.
A second, simpler frame comes from Raphael (1986), who taught children to ask where an answer comes from. Her four question-answer relationships are Right There, Think and Search, Author and You, and On My Own. The first two sit "in the book"; the last two sit "in my head". The two frames line up well, as the table below shows.
One short story to test every type
Every example in this article is written on the same story. It is original, 80 words long, and pitched at CEFR A2. You may use it in class.
The Red Bicycle (A2, 80 words)
Sam wanted a red bicycle more than anything. In March he started washing cars every Saturday, and he put the money in a jar. In June his friend Leo laughed at him. 'Why work so hard? Just ask your parents!' Sam said nothing and kept washing. In August the jar was full. He walked to the shop and bought the red bicycle himself. When Leo saw it, he stopped laughing. 'Can you show me how you did it?' he asked.
The moral is small and clear: patient work gets you what you want, and it earns respect. It gives us the evaluation question.
The six types, with an example question on the story
| Type | What it checks | Example question | Raphael's label |
|---|---|---|---|
| Literal | A fact stated in the text | What did Sam do every Saturday? | Right There |
| Reorganisation | Facts joined from two places in the text | How many months did Sam wash cars before the jar was full? | Think and Search |
| Inference | Something implied but not stated | Why did Leo stop laughing? | Author and You |
| Prediction | What might happen next, supported by the text | What do you think Leo will do next Saturday? | Author and You |
| Evaluation | A judgement about the text as a whole | What is the moral of the story? Do you agree with it? | Author and You |
| Personal response | The reader's own feelings, tied to the text | Have you ever saved for something you wanted? How did it feel? | On My Own |
Literal
Answer: he washed cars. Question and answer share one sentence. Day and Park (2005) note that teachers usually check literal understanding first.
Reorganisation
Answer: five months, from March to August. The reader must find two dates and count. Day and Park (2005) report that their students find these questions somewhat harder than literal ones.
Inference
The text says Leo stopped laughing; it does not say why. The reader joins the text to common sense: Leo saw that the plan worked. The answer is right or wrong, but the proof is indirect.
Prediction
Day and Park (2005) separate two kinds. While-reading prediction can be checked by reading on. Post-reading prediction, as here, has no single right answer, but every answer must be supported by the text. "Leo will start washing cars" is supported; "Leo will move to Spain" is not.
Evaluation and personal response
The moral question asks for a judgement about the whole story, so it counts as evaluation. Personal response asks for feelings, but Day and Park (2005) insist the answer must still relate to the story. For prompts that draw out a moral, see discussion questions to draw the moral from a story.
Which types are safe for A1 and A2 readers?
Literal and reorganisation questions are the safe core for beginners. The Council of Europe (2020) describes an A1 reader as someone who understands very short, simple texts a single phrase at a time, rereading as required. An A2 reader understands short, simple texts on familiar, concrete matters (see CEFR levels A1 to C2 explained). Both descriptors are about finding stated meaning, which is what literal questions test.
Inference and evaluation are not banned at A2, but they need help. Day and Park (2005) say that beginning and intermediate readers often read word by word and are reluctant to go beyond the literal meaning. Their advice is to ask how and why questions anyway, and to model possible answers. Raphael (1986) reports that the youngest pupils respond best to a two-way split, in the book or in my head, before they learn four categories.
- Keep the text in front of the reader. Day and Park (2005) are blunt: the aim is comprehension, not memory.
- Ask inference and evaluation questions in the simplest form: a yes or no question with a why follow-up.
- Give four options in multiple choice. Day and Park (2005) found that this works best for low-proficiency students.
- Avoid tricky wording and negatives; they test reading of the question, not of the story.
- Let the moral question be the one open question, and accept short answers.
The forms a question can take
Type and form are separate choices. The type is what you check; the form is how you ask. Day and Park (2005) list five forms: yes or no, alternative (this or that), true or false, wh- questions, and multiple choice. Here is the literal question on the story in each form.
- Yes or no: Did Sam wash cars on Saturdays?
- Alternative: Did Sam wash cars or sell cakes?
- True or false: Sam washed cars every Sunday. (False.)
- Short answer, a wh- question: What did Sam do every Saturday?
- Multiple choice: What did Sam do every Saturday? (a) washed cars (b) sold cakes (c) cleaned windows (d) walked dogs
Each form has a cost. Yes or no, alternative and true or false give a 50 per cent chance of guessing, so Day and Park (2005) recommend a follow-up question. True or false items are hard to write well, because the false statement must catch a real misunderstanding. Multiple choice is easiest to mark and suits literal questions best.
Vocabulary questions are a special case of the literal type and deserve their own care; see how to write vocabulary-in-context questions.
How to mix types in one quiz
Grabe and Stoller (2020) describe reading for general comprehension as grasping the main ideas and relating them to what you know, not remembering most of the details. A quiz that asks only for details therefore tests memory more than comprehension. Day and Park (2005) cite classroom research in which pupils did best on factual recall because that was what their teachers asked most.
A working shape for a short story quiz of six to eight questions:
- Two or three literal questions on the main events, in the order they happen.
- One reorganisation question that joins two parts of the story.
- One or two inference questions on a character's reason or feeling.
- One prediction or one evaluation question, and make the evaluation question the moral.
- One personal response question, if the quiz is for discussion rather than a score.
Raphael (1986) gives a similar shape. During reading, Think and Search questions should dominate and build towards Author and You; after reading, use Author and You and On My Own. Keep the quiz short. Day and Park (2005) recommend no more than ten questions for a text of about 600 words, and warn of "death by comprehension questions". See How many questions should a reading quiz have?.
How iRead handles this
iRead runs a quiz after every session, and practice questions are available between sessions. Questions are generated from the story itself: real sentences from the story, the grammar that appears in it, comprehension of the plot, and the moral of the story. Every story and quiz is levelled to the CEFR, from A1 to C2. Teachers see retention per word, per game and per class.
The core principle: Decide what each question checks before you decide how to ask it. A quiz is balanced when a pupil who finds the facts but misses the point cannot score full marks.
See it in iRead: Every iRead quiz is built from the story the class has just read: its sentences, its grammar, its plot and its moral. Each one is levelled to the CEFR.
See how iRead worksKey takeaway
Six question types, one story, one balanced quiz. Start with literal questions, add one that joins the text together, one that reads between the lines, and one on the moral.
Frequently asked questions
Can A1 readers answer inference questions?
Some can, if the inference is small and the form is simple. Ask a yes or no question about a character's feeling, then ask why. Keep the text in front of them. Day and Park (2005) found that beginners are reluctant to go beyond the literal, so model an answer first.
Is a question about the moral an evaluation question?
Yes. Asking what a story teaches needs a judgement about the whole text, which is how Day and Park (2005) define evaluation. Add "do you agree?" and it also invites a personal response. Accept any answer that the story supports.
Should every question in a reading quiz be multiple choice?
No. Multiple choice is quick to mark and suits literal questions, but pupils can guess. Mix in true or false with a follow-up, and one short-answer question. Day and Park (2005) suggest four options for low-level learners, with wrong answers that look plausible.
How do I know which type a question is?
Ask where the answer comes from. In one sentence of the text: literal. In two places: reorganisation. Implied but not stated: inference. After the story ends: prediction. A judgement: evaluation. The reader's own life: personal response. Raphael (1986) taught children this same check.
Sources
- Day, R. R., & Park, J.-s. (2005). Developing reading comprehension questions. Reading in a Foreign Language, 17(1), 60-73. doi.org/10.64152/10125/66599
- Raphael, T. E. (1986). Teaching question answer relationships, revisited. The Reading Teacher, 39(6), 516-522. www.jstor.org/stable/20199149
- Alderson, J. C. (2000). Assessing Reading. Cambridge University Press. doi.org/10.1017/CBO9780511732935
- Nuttall, C. (2005). Teaching Reading Skills in a Foreign Language (3rd ed.). Macmillan Education. www.macmillanenglish.com/us/catalogue/methodology/teaching-reading-skills-in-a-foreign-language
- Grabe, W., & Stoller, F. L. (2020). Teaching and Researching Reading (3rd ed.). Routledge. doi.org/10.4324/9781315726274
- Council of Europe (2020). Common European Framework of Reference for Languages: Learning, teaching, assessment. Companion volume. Council of Europe Publishing, Strasbourg. rm.coe.int/common-european-framework-of-reference-for-languages-learning-teaching/16809ea0d4