What happens when the teacher stops supplying every question, but still teaches pupils how to ask? A practical account of question generation, improvement, prioritisation and follow-through for Singapore classrooms and families.
A curious student looks at a photograph of rainwater pooling outside a school and asks, “Why is there so much water?” Another asks whether the drain is blocked. A third wonders whether a heavy downpour and a clogged grate could produce the same puddle. Which child has asked the best question? It depends on what they need to learn next. A wonderful-sounding question is not automatically answerable, and a narrow question is not automatically shallow. What matters is whether the student can make uncertainty into a manageable inquiry, choose evidence and recognise what the evidence does not settle.
The Question Formulation Technique, usually abbreviated QFT, is a teachable method for doing that work. Developed by the Right Question Institute, it deliberately separates producing many questions from evaluating and improving them, before learners choose priorities and decide how to use the results. For parents in Clementi wondering why a child can answer worksheets yet rarely asks a useful follow-up, the mechanism is important: pupils can be taught to design a better question instead of waiting for an adult to deliver one. The method is not a replacement for subject knowledge or teacher explanation. It is a bridge between noticing a puzzle and organising a responsible response.
The 50-second answer
QFT teaches students to generate, transform, select and use their own questions. The teacher offers a brief statement, image, object or situation that provokes inquiry without already telling learners what to ask. Pupils collect questions without immediately debating, judging or answering them. Then they examine which questions call for short facts and which invite explanation, practise changing one kind into the other, and choose a few worth investigating for a stated purpose. They make a next-step plan and reflect on how their questioning changed. The value is not a longer list; it is stronger ownership of inquiry combined with a route into evidence. It fails when the focus is vague, the activity becomes a popularity contest, the teacher confuses enthusiasm with understanding, or the class never follows up on the chosen questions. Use it to begin a topic, prepare research or diagnose what remains uncertain—not in place of explicit teaching of essential facts.
The shortest usable version: Put up a photograph of a rain gauge beside a flooded pathway. Ask children to record questions first, not answers. Compare “Did it rain yesterday?” with “What measurements could distinguish heavy rain from poor drainage?” Decide which is useful for a five-minute investigation and which needs a longer project. Then obtain suitable data and test what can actually be concluded.
The overlooked gap between being interested and knowing how to inquire
Classroom curiosity is often treated as a personality trait: some pupils supposedly have it, others do not. That view overlooks a skill problem. A learner may be eager to understand a news headline but struggle to turn the excitement into a checkable question. An anxious pupil may hide a brilliant uncertainty because the first draft sounds clumsy. Another may ask so many questions that a lesson ends without a single one being investigated. These are different difficulties. They call for different teaching moves.
The purpose of a structured question-generation routine is not to make everyone talk more. It is to give all pupils a low-risk chance to notice a gap, write the first imperfect wording and then refine it. Separating invention from evaluation is especially useful here. In ordinary conversations, an adult sometimes answers the first question immediately. That can be kind and efficient, but it also closes the space for other questions. In QFT, an initial burst is protected from premature answering so the learner can see alternatives before committing to a path.
Imagine two Secondary 1 Mathematics students facing a graph of daily temperatures. The first asks, “What is the highest value?” The second asks, “Would a different vertical scale change how dramatic the rise looks?” Neither question is inherently bad. One checks reading accuracy; the other inspects representation. A useful teacher does not rank them purely by cleverness. Instead, the teacher asks what decision or claim each question could support. This is how a list of curiosities becomes an inquiry plan.
A crucial distinction: students are generating questions, not inventing facts
The routine begins with agency, but it must end with epistemic responsibility. A student may generate “Do plants grow faster in darkness?” That is a legitimate thing to ask. It is not evidence that plants generally do so. The distinction is basic and often lost when inquiry is celebrated as an alternative to instruction. Young people need vocabulary, conceptual models and evidence standards to investigate responsibly. They should be free to ask a mistaken question without being taught that every premise in a question is correct.
Teachers can model the difference between three forms: a question that asks for a fact, a question that asks for a mechanism, and a question that contains an assumption requiring repair. “When was the bridge opened?” requires a date. “Why does the bridge need expansion joints?” requires a physical explanation. “Why do expansion joints make bridges weaker?” presupposes a consequence that may not be true. Before investigating the third, a learner needs to ask whether joints weaken the relevant structure under the specified conditions. This is inquiry with intellectual brakes, not unrestricted speculation.
A good question may still lead to a small answer. If the class needs to know which material conducts electricity, a short factual investigation may be exactly right. The test is not whether a sentence starts with why instead of what. It is whether the question matches the learning purpose and allows a defensible answer.
What the teacher gives first: a focus that creates space
The method’s central design challenge is the Question Focus, often called the QFocus. It is a stimulus around which learners develop questions. The focus is not the same as a teacher’s leading question. A map with two neighbourhoods showing different shade cover may work better than “Why should the council plant more trees?” because the latter already suggests a conclusion. A short statement such as “Two methods gave different answers to the same problem” may provoke comparison, measurement and error-checking without telling the class which explanation to prefer.
A Clementi Primary 5 Science class could see two photographs of identical-looking ice cubes placed on a metal tray and a wooden board. Their task at first is to generate questions, not announce that metal is colder. One pupil asks whether both surfaces started at the same temperature. Another asks whether the tray conducts heat differently. A third asks whether the images were taken at the same time. Each question reveals a possible confound or explanation. The focus works because something matters and enough context is visible to make inquiry possible.
A weak focus can be too abstract. “Science” leaves pupils unsure where to begin. A loaded focus can secretly force an answer. “Why is online homework always better?” frames a contested claim as fact. A focus can also be too obscure. A diagram filled with unfamiliar labels may produce questions mainly about the labels, not the intended mechanism. The teacher should try the focus personally: can it plausibly produce both simple and substantive questions from students at this level without providing misinformation?
How the thinking changes as a question travels
The first movement is divergence. The learner seeks possibilities and is not yet obliged to commit to one. Asking for immediate evaluation at this stage may privilege quick speakers and punish tentative ideas. A list of ten imperfect questions can show more of the conceptual landscape than one polished question chosen too early. This does not mean every question is equally good. It means the judgement is deliberately postponed so the pool of possibilities has room to form.
The second movement is discrimination. Students notice that some questions can be answered from a single reliable source or measurement, while others require comparing explanations, designing observations or deciding what counts as evidence. Turning “Does the plant need light?” into “How does changing light availability affect growth over a week?” alters the job ahead. The first is a potential fact check; the second is a possible investigation, but only once plant type, water, temperature and measurement are specified.
The third movement is selection. A class cannot investigate everything. It chooses questions according to the learning goal, time, resources, ethics and likely evidence. The fourth is action: a suitable reading, measurement, interview, calculation or discussion occurs. The fifth is reflection: learners examine which questions became stronger, what they found and what remains unresolved. Every stage has a different success criterion. Counting questions in stage one is not an adequate measure of the quality of stage five.
An actual 45-minute Clementi classroom route
Start with a concrete focus: a simple, clearly labelled drawing of a school garden in which plants close to a covered walkway look smaller than plants in an open bed. Tell students that the picture is a teaching illustration rather than a report of actual school data. Do not state the cause. Explain that the lesson goal is to produce a manageable investigation question about plant growth, not prove that shade is harmful.
Minutes 0–5: Orientation. Give pupils the focus and define necessary terms: growth can mean height, leaf area, biomass or another measure; light is not the only variable. Ask them to distinguish a drawn observation from a measured result. This protects the lesson from pretending its picture is evidence.
Minutes 5–12: Question collection. Pupils write questions without commentary. A class might produce: Were all seedlings planted at the same time? How much sun does each group get? Is there a water pipe near the open bed? Could a different type of plant be present? Does the covered bed have shallower soil? Can we measure growth over ten days? What would happen if the pots changed position? Encourage exact transcription, including awkward wording, so pupils can later see what they improved.
Minutes 12–20: Inspect and reshape. Students identify questions that invite a fixed piece of information and questions that ask for more extended explanation or testing. They deliberately revise one in each direction. “Is there shade?” becomes “How much does daily light exposure differ between the beds?” “How does the location affect the plants?” becomes “Were the seedlings watered on the same schedule?” The comparison teaches that open and closed forms are tools for different jobs.
Minutes 20–28: Choose. Tell pairs the available resources: two thermometers, a light meter if the school has one, growth charts, one short reading and twenty minutes in a later lesson. Each pair identifies a question that fits the resources, a question requiring more time and a question that is ethically or practically unsuitable to test. The most dramatic idea is not necessarily the best project.
Minutes 28–38: Plan. Pupils record the selected question, what data would be needed, what factors should be held constant and what finding would not establish causation. If the class cannot perform a real experiment, it may compare supplied datasets or critique the design of a proposed experiment. Do not promise a scientific outcome merely because pupils are excited.
Minutes 38–45: Reflect and assess. Each pupil writes the original question, its revision, why the revision is more usable and one thing they still need to learn before answering. The teacher assesses clarity, alignment with evidence and awareness of controls. A child who asks only two thoughtful questions may demonstrate more progress than one who generates twenty unrelated questions.
The Singapore Primary Science example: from observation to fair test
Suppose a Primary 4 child sees droplets on the outside of a cold bottle. The first question is “Why is the bottle leaking?” That is understandable: a wet surface often comes from water escaping. A teacher using question formulation need not embarrass the student or immediately explain condensation. The class can generate alternatives: Is there a hole? Could the droplets come from surrounding air? Is the drink colder than the classroom? Does it happen with an empty sealed bottle? What if the outer surface is dried and observed again?
The quality of the questioning improves as students identify discriminating observations. “Is there a hole?” is checkable. “Would a sealed empty cold container develop droplets?” begins to separate the leak hypothesis from condensation. The key teaching move comes next: pupils must learn the science of water vapour, cooling and condensation. Without that explicit knowledge, they may ask clever questions yet leave with the false belief that cold objects produce water. The routine creates a route into content; the content explains the phenomenon.
For a PSLE-style response, a teacher can then ask each pupil to write an independent explanation based on the correct scientific model. The discussion question and the examination explanation serve different purposes. One is diagnostic and exploratory. The other is an accurate account of a mechanism. A mature lesson connects them instead of claiming that question generation alone is sufficient assessment preparation.
The Mathematics example: questions that expose hidden assumptions
A Secondary 1 student sees two shopping offers: a 20% discount from one price and a fixed dollar discount from another. A useful focus contains the actual prices, not only bright promotional words. Students may ask: Which saves more money? Does the comparison change with the original price? Is there a minimum spend? Are we comparing equal quantities? Is service tax applied before or after the reduction? Is the discount capped? These questions do not all belong to the same domain: some concern arithmetic, others reading conditions and others the fairness of the comparison.
The teacher invites a revision. “Which offer is better?” is too broad until better is defined. “For a S$60 item, which offer leads to a lower final price before any additional fees?” is testable. Now the maths can happen: one offer gives 20% off, producing S$48. Another gives S$15 off, producing S$45. On those assumptions, the second is cheaper. But if the comparison item, eligibility or fees change, the conclusion may change. Better questions lead to cleaner arithmetic because the quantities and conditions become explicit.
A useful extension is to let learners invent a case where the ranking reverses. The teacher might compare a percentage discount and a fixed discount over varying prices. At what original price do the savings match? If the fixed discount is S$15 and the percentage reduction is 20%, then 0.20p = 15, giving p = 75. This threshold question connects everyday wording to an equation. The method was not a substitute for algebra; it made the algebra’s purpose visible.
The English example: comprehension beyond collecting question words
A Primary 6 student reads a short story about a character who refuses to join a school performance. A weak response is to ask pupils to produce three questions and reward any three. The exercise can easily become a list of shallow who-when-where items. Instead, provide a short excerpt and ask what information the reader needs to understand the refusal. Questions might concern an earlier promise, fear of embarrassment, another character’s reaction or a narrator’s omission. Students then classify which answers are explicitly stated and which would require inference or later textual evidence.
This is where a canonical boundary matters. The separate Question–Answer Relationships work asks where an answer can be found or constructed; question formulation asks how pupils originate, reshape and prioritise the questions before answering. These jobs can cooperate, but they are not interchangeable. One does not automatically teach the other.
The follow-through is a written answer with text support. Suppose the chosen question is “Is the refusal more about fear of failure or loyalty to another friend?” Students should quote or paraphrase relevant details, explain what those details permit and acknowledge when the excerpt cannot settle the issue. A question that stimulates a hundred guesses but no return to the text has failed the comprehension purpose. The best questions turn reading into a more disciplined search for evidence.
The Humanities example: deciding what a source can tell us
A class looks at a historical poster encouraging citizens to conserve a scarce resource. Their first questions could include: Who produced it? When? Who was the intended audience? Which behaviour did the poster want to change? Does the design show that people actually followed the message? What other records would reveal public response? There is a qualitative change from decoding an image to evaluating its provenance and limits.
The teacher should be explicit: a source’s existence does not prove its success. A campaign poster tells us that an organiser wanted to persuade an audience, not necessarily that the audience was persuaded. Students can use question generation to discover this difference, but they still need subject instruction in sourcing, contextualisation and corroboration. The most valuable questions are sometimes the ones that reveal what a single source cannot establish.
For an individual exit response, ask students to identify one question the poster can answer, one that requires corroboration and one that cannot be answered reliably with the available material. This makes source literacy visible and prevents a class from treating every question as answerable by a quick search. It also keeps the activity connected to Humanities assessment, where inference must rest on evidence rather than speculation.
How parents can use the method without turning dinner into an oral examination
At home, a parent may reasonably want a direct answer: “How many more questions must we ask when homework is already difficult?” The answer is not to run a full classroom protocol every evening. A small version can start with something the child genuinely notices: a cooking measurement, a plant on the balcony, a misleading chart in the news or an unfamiliar word in a story. Invite a few questions without correcting the first one immediately. Then ask which question can be answered from what is already available and which would require a book, experiment or trusted source.
Avoid using the routine as a test of cleverness. When a child asks “Why is that?” five times, the adult might feel challenged or tired. Instead of evaluating the child’s character, help narrow the referent: “Which part surprised you—the sound, the timing or the result?” Ask the child to pick one manageable inquiry and decide how to check. An answer supplied by an adult is sometimes the appropriate endpoint, especially when safety or factual accuracy matters. Agency does not mean withholding helpful information.
A useful home record has three lines: “I wondered…”, “I changed my question to…”, and “I found out / still cannot tell…”. One careful example each week is more informative than a long list of invented inquiries that nobody revisits. If the student is younger or developing English proficiency, oral questions, sketches or mixed-language notes can serve the initial generation stage, followed by explicit work on precise academic wording.
The teacher’s hardest craft: choosing the degree of openness
There is an interesting tension at the heart of QFT. A very open focus can stimulate imagination, yet leave novice learners without enough knowledge to discriminate promising ideas. A highly guided focus can support accurate reasoning, yet reduce question generation to a hunt for the teacher’s preferred response. Good design is not a permanent commitment to one end of that spectrum. It changes with the learners’ preparation and the task.
Before a new topic, questions may be diagnostic: what do students wonder or mistakenly assume? After an explanatory lesson, the focus may invite transfer: which conditions would make this model fail? At the end of a unit, questions may help select a research extension or compare competing sources. A teacher should be able to say why this moment benefits from student-generated questions rather than a worked example, explicit modelling, practice, retrieval or direct feedback. If the aim is learning a core procedure that no one has yet been taught, letting the class guess indefinitely is not a virtue.
The focus may also reveal bias. A photograph that shows only one side of a conflict, a graph with a truncated axis or a deliberately sensational statement can steer the questions toward a false premise. Sometimes such bias is itself the object of critical media literacy; if so, teach pupils to identify it. Otherwise, balance the materials before starting. QFT is not a licence to launder misleading source material through student curiosity.
Why separating creation from criticism can help
When students know every tentative question will be judged instantly, some self-censor. They may choose safe questions whose answers they already know, particularly in a competitive class. A temporary no-evaluation period permits more varied starting points. That separation resembles a general design principle used in brainstorming: divergent production before convergent selection. It does not imply that producing more ideas always produces better understanding or that quantity alone should earn marks.
The educational mechanism becomes sharper when learners later revise the questions. A question changes not only in wording but also in the test it demands. “Can we use more water?” might become “What change in water use would be needed to produce a measurable increase in plant growth under otherwise comparable conditions?” The latter may still need tightening; that is fine. The learner can now see what data would be required and which variables matter. The revision is evidence of conceptual progress even before the answer arrives.
This also explains why a facilitator should not correct every question during generation. If adults dominate the process, the final questions may be elegant but no longer student-owned. The adult can still offer vocabulary, access support and later feedback. The aim is a temporary transfer of initiative, not an abdication of professional responsibility.
Closed questions are not inferior to open questions
A common mistake in inquiry lessons is to label open-ended questions “deep” and closed questions “shallow.” That is too simple. “Was the sensor calibrated before the measurement?” may require a yes or no, yet it can be decisive for judging whether a dataset is trustworthy. “Why is the world unfair?” is open-ended but too broad to investigate as a single classroom inquiry. Question form changes the information task; it does not independently guarantee intellectual quality.
Teachers can show this with paired revisions. “Did the character apologise?” can become “How does the timing of the apology change the reader’s view of the character?” The first locates an event; the second interprets consequences. But the reverse revision can be just as useful. “Why does the character seem dishonest?” could become “Which two statements by the character directly contradict each other?” The narrower question creates a factual foundation before interpretation.
For Science, “Does the object float?” becomes “Which measured properties help predict whether it floats in this liquid?” Conversely, “What factors explain every floating and sinking phenomenon?” becomes “Does this object float in fresh water at room temperature?” The best pair may be one factual check and one explanatory question. Pupils need both in a responsible investigation.
How to choose a priority without mistaking popularity for quality
After question generation, a classroom can become a voting exercise: the most amusing or dramatic proposal wins. That may sustain enthusiasm but not necessarily learning. Priority should follow stated constraints. The class might choose a question that helps meet the curriculum objective, can be investigated safely, fits the available time, requires evidence pupils can obtain and leaves an opportunity to explain a mechanism.
A simple scoring conversation can use four plain criteria without turning the activity into a bureaucratic spreadsheet: relevance, answerability, evidence and learning value. Relevance asks whether the question addresses the focus. Answerability asks whether it can be bounded. Evidence asks what the learner could consult or measure. Learning value asks whether the investigation would reveal something worth understanding. Students can rank differently for different purposes. A question may be excellent for a semester project and poor for a twenty-minute lesson.
When comparing priorities, ask learners to defend their selection with one trade-off. “We chose the drainage question rather than the weather question because the available rainfall record is already known, but the blocked-drain claim requires an observation we can conduct.” This sentence demonstrates planning. “Because it sounds more interesting” may be an honest preference, but it is not sufficient when the assignment is evidence-based investigation.
Follow-through is the difference between inquiry and decoration
One weakness in many questioning activities is that the most exciting part is the colourful sheet of questions on the wall. Students admire the sheet, the class moves on, and nobody returns to the questions. That is a missed opportunity. A good question is not merely a classroom artefact; it should direct an action. The next step may be reading a chapter, interviewing a suitable expert, comparing two representations, collecting a safe measurement, revising a model or checking whether a premise is wrong.
Every chosen question needs a work contract, however informal: what will we try, which resource will we use, when will we know enough to give a provisional answer, and what limitation will we record? A study group researching whether an infographic exaggerates a trend might obtain the original dataset, reproduce the graph with a clearly marked axis and write two interpretations. A pupil investigating how a metaphor changes tone might compare alternative word choices in context. The exact action depends on the domain, not on the generic attraction of inquiry.
This is where independent work matters. Each pupil should produce at least one personal explanation, evidence note or changed question, even when the activity was collaborative. Otherwise, a talkative partner may create an impressive collective product while a quieter partner remains unsure what happened. Group curiosity is not automatically individual knowledge.
A diagnostic case: the class asks lots of questions but learns little
Imagine a Secondary 2 lesson with a spectacular demonstration. Pupils write forty questions. The teacher praises their creativity and closes the activity. A week later, many students cannot explain the underlying Science concept. What went wrong? The questioning event achieved divergence but omitted selection, investigation, explicit explanation and retrieval. It may have increased interest without securing learning. The problem is not that questions were bad; it is that the learning sequence stopped midway.
A repair would be to select two or three questions that connect directly to a causal model. The teacher teaches relevant vocabulary, demonstrates the mechanism or guides a fair test, asks pupils to answer the chosen questions with evidence, and checks learning later with a new example. Students might be asked which original question has now been answered and which requires additional information. That closes the loop without pretending all uncertainty has disappeared.
An equally common failure is the reverse: the teacher brings ready-made answers, then asks pupils to generate questions that lead exactly to those answers. In that case, the routine becomes performative. Pupils learn to guess what the adult wants rather than investigate. A better teacher can accept productive questions that were not anticipated while maintaining the subject’s accuracy and the lesson’s boundaries.
The difference between measuring engagement and measuring understanding
Questions are visible. A teacher can count them, photograph the board and hear lively conversations. That visibility is attractive, but it can mislead. A rise from four questions to twenty may reflect improved comfort or fluency rather than stronger comprehension. Conversely, an expert learner may ask only one sharply framed question because they already understand the surrounding material.
A more meaningful evaluation uses multiple indicators. First, does the student distinguish the information being sought? Second, can they repair an ambiguous or assumption-loaded question? Third, can they select evidence that would address the question? Fourth, can they answer accurately and name a limitation? Fifth, can they reuse these habits with an unfamiliar example several weeks later? These outcomes should be interpreted alongside actual subject knowledge.
A teacher can collect three examples before and after an inquiry unit: an initial question, a revised question and a post-investigation explanation. The difference between the first two indicates question-quality development; the final explanation indicates whether the work yielded learning. A student who writes an elegant research question but misunderstands the experiment still needs teaching. A pupil who understands the content but asks ordinary questions may need an extension rather than remediation.
When a question is too large, too small or impossible to test
A question is too large when it contains several investigations disguised as one. “How does technology change learning everywhere?” could become a year’s programme. For a lesson, ask whether a specified feedback feature helps a particular type of revision, and decide what evidence could inform the comparison. A question is too small when it can be answered instantly but adds little to the purpose. “Is the title blue?” may still be useful in visual analysis, but not as the sole question in a unit about argument quality.
Some questions cannot be answered with the resources available. That is not a failure of the child. “Would this policy make every pupil happier?” lacks a clear measure and probably asks for an impossible universal conclusion. The useful revision may be “What outcomes would show that more pupils feel able to ask for help, and how could we collect them ethically?” Learners should be allowed to keep a parking list of worthwhile questions that exceed the present investigation’s scope.
Other questions are unsuitable because testing them would be unsafe, invasive or unfair. A child should not be asked to collect classmates’ private family information to satisfy curiosity. A science demonstration should not invite uncontrolled experimentation with hazardous materials. A question about personal identities or dignity should not become a classroom debate over whether some classmates deserve respect. Curiosity is valuable within ethical constraints.
English-language support without deciding the questions for students
Students with a narrower English vocabulary may notice the same meaningful puzzles as fluent peers but struggle to express them. The teacher can provide stems such as “What would happen if…?”, “How could we tell whether…?”, “What is different between…?” and “Which evidence would show…?” They are temporary linguistic scaffolds, not answers. Visual supports, bilingual planning notes and rehearsed oral questions may help students participate before their writing becomes fully precise.
Avoid treating grammatically imperfect wording as lack of thought. The question “Why plant small there?” can contain an excellent inquiry about environmental conditions. The teacher can first recognise the intended comparison, then help reformulate: “Which differences between the two places might explain the plants’ different heights?” The pupil owns the puzzle; the adult lends language that makes the question investigable.
At the same time, language precision matters. “Why does it change?” may refer to temperature, colour, speed or a result in a graph. Teaching pupils to name the variable and context is more useful than praising vagueness as creativity. The aspiration is access with improving accuracy, not permanent reliance on prompts.
Question formulation is not the same as asking a teacher for help
When a pupil is stuck on homework, the immediate need may be a hint, a worked example or correction. A question-generation routine can help identify the obstacle, but it should not become an excuse to refuse instruction. “What part is confusing?” may be too broad for a distressed learner. “Can you locate the first line where the working stops making sense?” is often kinder and more diagnostic. After identifying the obstacle, the adult may teach directly.
The skill is most valuable when the learner can distinguish a request for an answer, a request for an explanation, and a question that can guide independent investigation. All three are legitimate. A child who asks “Is this correct?” can learn to add “I think the error may be where I combined unlike terms; can you check that step?” That turns helplessness into a more specific request without pretending the student must solve everything alone.
For older learners, a question log can reveal recurring patterns. Perhaps the pupil repeatedly asks for reassurance about final answers but seldom asks why a method applies. Perhaps they ask excellent conceptual questions but avoid routine practice. The teacher can respond accordingly. The goal is not maximum independence at all costs; it is appropriate agency supported by timely expertise.
How it differs from three neighbouring approaches
Hinge questions are prepared by the teacher to decide what should happen next in a lesson. They target a known misconception or knowledge checkpoint. QFT puts the formulation of the questions into students’ hands and then teaches them to refine what they generated. Both reveal information, but the direction of question ownership differs. Teachers can use a hinge question after a QFT investigation to determine whether the intended concept was learned.
Question–Answer Relationships help readers reason about where an answer is located: stated in the text, assembled from several textual details, inferred using knowledge or answered through the reader’s own reasoning. QFT is upstream: it helps learners invent and choose a worthwhile inquiry. A comprehension lesson can use both, but each has a separate canonical job and should not be relabelled as the other.
Socratic questioning often involves a teacher or interlocutor probing assumptions through successive questions. It can be intellectually powerful, but the question path may remain controlled by the experienced questioner. QFT makes the ability to formulate questions itself the explicit learning target. This is a matter of emphasis, not a claim that one method is universally better. Sometimes direct exposition is faster and more effective; sometimes inquiry deserves extra time.
What current evidence permits us to say
The Right Question Institute explains the design of QFT and maintains a research programme examining question formulation in education and other settings. Harvard Graduate School of Education offers professional learning led with the method’s originators. Those facts establish an identifiable intervention, not proof that every adoption yields large learning gains. Reports from users and professional-learning descriptions are useful for implementation ideas but should not be treated as independent causal evidence.
Research on student-generated questions is encouraging but more specific than broad promotional claims. For example, Taboada and Guthrie examined links among questioning, prior knowledge and reading comprehension in a sample of 360 third- and fourth-grade pupils. Their findings support an association between students’ questioning and comprehension after accounting for prior knowledge; that does not by itself prove that this particular classroom protocol causes the same outcome. A 2022 study by Maplethorpe and colleagues investigated associations involving question generation and literacy measures in a sample of 89 fifth- and sixth-grade pupils. Such findings suggest questions matter, but they do not justify a universal effect size for QFT.
The Right Question Institute also reports classroom studies and a 2025 survey of 394 educators, conducted from 20 March to 10 April 2025. A survey of practitioners can describe their experience, but without random assignment or other careful comparison it cannot establish the gain caused by the technique rather than teacher enthusiasm, prior skills or contextual differences. The strongest responsible conclusion is conditional: explicit practice in question formulation is plausible and promising, but its effects depend on the focus, subsequent teaching, evidence use, age and implementation.
The assessment trap: rewarding the question that sounds impressive
A rubric that asks for “interesting questions” risks bias toward elaborate vocabulary and unusual topics. Some children perform originality; others state a decisive but plain question. An inquiry rubric should prefer clarity, relevance, evidence alignment and revision. A student who changes “Why is the plant unhappy?” to “How does the amount of light differ between these two locations?” has made a useful intellectual move even if the words are simple.
For more advanced students, add assumption checking. Does the question already contain the answer the student hopes to find? “Why did the new method fail?” presupposes failure; a more neutral version asks what evidence shows how the method performed. Another criterion is feasibility: could the available source or experiment actually address the question? A beautifully stated impossible test should not receive full marks for scientific planning.
Keep evaluation proportionate. The activity is a learning opportunity, not necessarily a high-stakes examination. A short exit note and one independent question revision can reveal more about pupil progress than a page of generic praise. When grades are necessary, disclose criteria before assessing and provide at least one worked example of improvement.
A model progression across Primary 3, Primary 6 and Secondary 2
For Primary 3, use tangible objects or clear images. A teacher might display two magnets of different shapes and invite questions about what each attracts. Children use simple wording and receive help separating testable properties from fanciful explanations. The follow-up can be one safe teacher-led demonstration with accurate language about magnetic materials.
For Primary 6, connect the technique to evidence-based written responses. Show a short non-fiction passage and a simple chart about recycling rates. Pupils generate questions, decide which the data can answer, and practise avoiding claims about motives that the chart does not contain. Then they write a summary that distinguishes reported rates from proposed explanations. The teacher explicitly models inference boundaries.
For Secondary 2, ask students to compare two historical accounts of a public decision or two mathematical representations of the same relationship. Questions should address provenance, underlying assumptions, comparability and missing information. Learners choose a question and assemble a short evidence brief. Their final judgement states what would change it. In all three stages, the routine matures from noticing to critical inquiry without requiring the teacher to surrender responsibility for content.
A short template for a student’s own inquiry notebook
A notebook entry can be only one page. At the top, copy the focus in a sentence. Below it, write the unfiltered questions in the language that first occurs. Next, circle two candidates and note why one is factual and the other explanatory, or why both require different kinds of evidence. Then rewrite the chosen question to name the relevant object, condition and outcome. Record one source or action that could answer it, followed by a sentence saying what the source cannot establish.
Finish with two reflections: “The most useful change I made was…” and “The answer I can defend so far is…, because…”. That last phrase matters. An inquiry notebook without an answer route may become a collection of attractive puzzles. The answer can be provisional. A provisional conclusion is not a failure if uncertainty is honest, the evidence is appropriate and the next test is clear.
The notebook is also useful for parent–teacher communication. Instead of saying “My child does not ask questions,” a family can bring one example that shows where the process breaks: generating ideas, narrowing, finding evidence or writing the final explanation. Each break calls for a different response. This is more actionable than treating inquisitiveness as a fixed talent.
Failure modes worth anticipating before the first lesson
Failure one: the focus contains a false premise. Pupils become productive at asking questions around misinformation. Repair it by fact-checking the stimulus and teaching source awareness. If a misleading premise is intentionally included for media literacy, label the task as evaluating a claim rather than treating it as established reality.
Failure two: question counts become a competition. The loudest or fastest pupils flood the board. Repair it with private writing time, equal opportunity to contribute and later quality criteria. A small set of discriminating questions may be excellent.
Failure three: every question stays broad. The class never names an outcome or source. Repair it with domain-specific narrowing: which variable, period, population, text passage or comparison? Offer a worked example of revision while leaving ownership of pupils’ questions intact.
Failure four: the teacher immediately answers everything. Learners outsource the inquiry. Repair it by protecting a short generation interval, then answer strategically once the class has chosen a direction. Do not withhold urgent safety information or core instruction.
Failure five: the task becomes a search-engine race. Students select the question with the easiest snippet rather than the most illuminating evidence. Repair it by requiring an explanation of source reliability and what the selected source can and cannot establish.
Failure six: nothing happens after prioritisation. Questions decorate the wall but produce no inquiry. Repair it by reserving time and resources for follow-through and asking for an independent return product.
Failure seven: the teacher treats all claims as equal. Inaccurate premises go uncorrected in the name of student voice. Repair it with explicit knowledge teaching, evidence checks and clarity about settled facts versus contestable interpretations.
Failure eight: the method consumes an entire lesson that needed direct teaching. The routine is used because it looks engaging. Repair it by matching the tool to the instructional purpose. It is permissible to choose a three-minute prompt, a factual explanation or a worked example instead of the full process.
A parent route that fits into a real weekday
Imagine a Clementi family with homework, dinner and very little spare time. The child has read about an animal that appears in a city park. The parent says, “What do you wonder about it?” The child replies, “Why is it here?” Rather than immediately beginning an internet search, the parent invites two alternatives: “What food can it find here?” and “How could we tell whether it lives here or only passes through?” The family chooses one modest question and consults a reliable local wildlife resource, paying attention to whether the information concerns the same species and setting.
The entire exercise might be ten minutes. The crucial outcome is the child’s ability to say, “I first thought it had lost its home, but now I need evidence about habitat and food; seeing one animal does not prove why it came.” That is an achievement in reasoning even when the family cannot answer the ecological question fully. If the source remains uncertain, they leave the question open.
A parent can also use the approach when the child is worried about study performance. “Why am I bad at maths?” is an emotionally understandable but analytically unhelpful question. “Which step in ratio problems am I finding difficult, and can I get it right on three varied examples after feedback?” makes the next helpful action visible. The aim is not to turn feelings into data compulsively; it is to offer one manageable path from worry toward support.
A teacher route: teach the routine explicitly, then make it portable
The first time a class uses QFT, model the difference between a stimulus, a candidate question and a source of evidence. Reserve enough time for pupils to experience the full cycle. Explain that the initial no-judgement period is a temporary production rule, not an exemption from later accuracy. Agree on how students will contribute without interruption. Choose a focus appropriate to their knowledge and reading level, and prepare a realistic follow-up rather than hoping the best questions will magically find answers.
On subsequent lessons, the full protocol need not be repeated each time. A teacher can revisit one component: rewriting an assumption-loaded question, making an open question testable, identifying what a source can answer or improving the quality of a final reflection. These are portable skills. They should appear in Science investigations, Humanities sourcing, mathematics problem interpretation and English comprehension even when nobody uses the acronym.
Build a small evidence trail across a term: early question samples, post-feedback revisions, answered investigations and transfer to a new task. Discuss with colleagues whether gains are showing up in actual subject work. If pupils become more willing to ask but still cannot explain core content, increase content instruction and follow-through. The method should serve teaching rather than become an end in itself.
A learner route: take ownership of the next useful question
Students can apply the routine in an abbreviated personal form. Before opening a textbook, write what feels unclear without worrying about elegance. Then ask whether the uncertainty is about a word, a relationship, a method, evidence or an exception. Rewrite the question so an answer would change what you do next. Choose whether the best next step is a worked example, a specific page, a reliable source, an experiment or a teacher’s explanation.
After studying, answer the revised question without looking if that is appropriate, then check accuracy. If it remains unresolved, narrow further. “I don’t understand algebra” is a fog. “Why does multiplying both sides by the same non-zero value preserve the solutions of this equation?” is a learnable target. “What goes wrong if I multiply only one side?” creates an example that can test the explanation. Asking a useful question does not replace retrieval and practice, but it can tell the learner which practice will matter.
Be wary of the false pride of endless inquiry. There is no intellectual medal for refusing an explanation. Experts learn by asking, reading, calculating, testing and accepting correction. The skill is choosing the next step that can actually improve understanding.
FAQs: the answers parents and teachers usually need
Is the Question Formulation Technique just brainstorming?
No. Brainstorming resembles the early generation phase, but the full instructional mechanism also asks learners to examine question types, revise them, prioritise for a purpose, plan next steps and reflect. Without that convergence and follow-through, an idea list may not become an investigation.
Does every pupil need to generate lots of questions?
No. A larger initial pool can give learners something to select from, but question counts are not the learning goal. Assess whether questions become clearer, more relevant and better linked to evidence. Allow participation through writing or supported oral language.
Does asking open-ended questions automatically develop higher-order thinking?
No. Open-ended questions may invite explanation, but some are vague or impossible to investigate. Closed questions can protect data quality and establish essential facts. The cognitive demand depends on the reasoning and evidence required, not the grammatical form alone.
Is QFT proven to raise examination marks?
The wider literature supports paying attention to student questioning, and the method’s originators report promising studies and practitioner experience. But no single effect estimate from those sources establishes a guaranteed improvement for every Singapore examination. Content instruction, practice, feedback and task alignment remain essential.
At what age can pupils begin?
Young learners can generate oral questions about concrete objects and pictures. Older pupils can revise for precision, source quality and research design. The language and length should be adapted to developmental and curriculum needs.
Should the teacher refuse to answer questions during the routine?
Only the initial generation phase benefits from postponing routine evaluation. Safety issues, essential clarifications and direct teaching remain the teacher’s responsibility. Later phases explicitly require evidence and answers.
Can it help a learner who is afraid of being wrong?
It may reduce pressure by allowing a private first draft and delaying public judgement, but the result depends on classroom climate. Written options, teacher modelling and respectful feedback matter. It is not a clinical intervention for anxiety.
What is the difference between a question focus and a learning objective?
The learning objective states what students should know or be able to do. The focus is a stimulus used to invite student questions. A good lesson aligns them without simply placing the desired answer inside the stimulus.
Can parents run the full protocol at home?
They can, but usually a shortened cycle is enough: notice, ask, refine, select and check. The purpose is useful thinking, not creating one more homework assignment. Choose an everyday question with an accessible next step.
Does it fit Science and Mathematics as well as English?
Yes, provided the questions lead into disciplinary standards. Science needs variables, models and evidence; Mathematics needs definitions, quantities, examples and proofs; English needs text support and language accuracy. The method does not eliminate those demands.
What if the child’s question reveals an incorrect belief?
Record it respectfully, then investigate or teach the relevant knowledge. A mistaken premise is a valuable diagnostic signal but not a fact. Help the pupil rewrite the question without humiliation.
Is an unanswered question a bad outcome?
Not necessarily. A well-defined limit can be an important result: perhaps the source is insufficient or the proposed test cannot isolate the variable. Learners should still record what can be concluded and what evidence would be needed next.
How do I know whether the questioning lesson worked?
Look for a trace from initial wording through revision to justified answer or next action. Compare independent work on a new task. A colourful board of questions and an enthusiastic discussion are not sufficient evidence by themselves.
How is it different from a teacher’s hinge question?
A hinge question is designed by the teacher to reveal understanding at a decision point. QFT teaches students to originate and improve questions. Both have value, and they can appear in the same sequence for different reasons.
Could question generation waste time during intensive exam preparation?
Yes, if used without a clear purpose or follow-through. A short targeted version may help identify misunderstandings, but it should not displace necessary worked examples, timed practice, retrieval or feedback. Match the method to the learning need.
Why not simply search the internet for an answer?
Search can be a useful next step only once the learner knows what to search, which evidence counts and how to assess the result. A vague or assumption-loaded question produces unreliable results even when the search engine responds quickly.
The useful return: a better question changes the work
Imagine the same Clementi student returning to the rainwater photograph after a lesson. Their first question was “Why is it wet?” Their new question is “Given rainfall at this time, which observations would help distinguish a blocked drain from a temporary volume of water exceeding the drainage capacity?” That is not yet an answer. It is progress because the student can name competing mechanisms, identify the needed observations and recognise what one photograph cannot prove.
Now the teacher can teach, and the student can investigate. Read the relevant Science, inspect a safe dataset, define the measurement, compare explanations and write a cautious conclusion. The most valuable product of a question-formulation lesson is not the moment a child raises a hand. It is the later moment when the child knows which question to ask next—and what responsible work an answer would require.
Canonical scope and internal reading routes
This article owns the formal student-generated question cycle centred on a Question Focus: open production, improvement, purposeful prioritisation, follow-through and reflection. It does not own teacher-authored diagnostic questioning, reading question–answer source categories, all inquiry learning or every method for teaching curiosity. Relevant neighbouring routes include How Hinge Questions Work, How Formative Assessment Works, the How X Works hub, the How Education Works collection and the Clementi small-group tutorial reference.
Evidence, origin and further reading
- Right Question Institute, The Question Formulation Technique in Education and research overview: https://rightquestion.org/what-we-do/ and https://rightquestion.org/research/
- Dan Rothstein and Luz Santana, “The Right Questions,” Educational Leadership (ASCD), October 2014: https://www.ascd.org/el/articles/the-right-questions
- Harvard Graduate School of Education, Teaching Students to Ask Their Own Questions: Best Practices in the Question Formulation Technique: https://www.gse.harvard.edu/professional-education/program/teaching-students-ask-their-own-questions-best-practices-question
- Right Question Institute, guidance and usage FAQ: https://rightquestion.org/faq/
- Ana Taboada and John T. Guthrie (2006), “Contributions of Student Questioning and Prior Knowledge to Construction of Knowledge from Reading Information Text,” Journal of Literacy Research: https://doi.org/10.1207/s15548430jlr3801_1
- Lois Maplethorpe and colleagues (2022), “Student-Generated Questions in Literacy Education and Assessment,” Journal of Literacy Research: https://doi.org/10.1177/1086296X221076436
The Question Formulation Technique was created by the Right Question Institute (rightquestion.org). This explanatory article is independently written, and the illustrative Singapore classroom and family scenarios are constructed examples, not reported interventions or guarantees of outcomes. The evidence cited for student-generated questions and educator experience should not be represented as a controlled estimate for this specific implementation.