Questioning Techniques for One-to-One Tutoring: Beyond 'Do You Understand?'

A one-to-one lesson gives a tutor something no classroom teacher has: the ability to question a single student continuously, adaptively, and without an audience. Yet transcripts of real tuition sessions show most questions falling into two low-value types — 'do you understand?' (which students answer yes to regardless) and rapid-fire steps ('so what do we do next? and then? and then?') that test compliance rather than thinking. Better questioning is the cheapest upgrade available to any tutor, because it requires no new materials — only different habits.

Wait time: the three seconds that change everything

The classic research finding on questioning is embarrassingly simple: after asking a question, most teachers wait less than one second before rescuing the student with a hint, a rephrase, or the answer. Extending that pause to three to five seconds transforms the quality of responses — answers get longer, reasoning appears, and 'I don't know' declines. The one-to-one setting makes the silence feel more awkward, not less, because there is no class to hide it. Sit with it anyway. A useful private rule: after asking, count silently to five before saying anything, and if the student answers, count to three again before responding — the second pause often draws out the reasoning behind the answer, which is the part you actually needed to hear.

The funnelling trap

Funnelling is what happens when a tutor, faced with a stuck student, asks a sequence of ever-narrower questions until only one trivial step remains: 'What kind of triangle is this? So which theorem uses right angles? So what do we write first? So what is a squared plus b squared?' Each question gets answered, the problem gets solved, and the lesson feels productive — but the tutor did all the mathematical thinking and the student supplied arithmetic. The test for funnelling is to ask who chose the method. If the answer is always you, the student is practising obedience.

The alternative is focusing: questions that direct attention without dictating the step. 'What do you notice about this triangle?' rather than 'is it right-angled?'. 'What information have we not used yet?' rather than 'should we use the radius?'. Focusing questions feel riskier because the student might go somewhere unexpected — which is precisely their value, since where the student goes is diagnostic information.

Socratic prompts worth memorising

A small bank of reusable prompts covers most situations. The point is not the exact wording but the pattern: each prompt hands the thinking back.

• Before starting: 'What kind of problem is this? What made you say that?'
• When stuck: 'What do you know? What are you trying to find? What connects them?'
• After a step: 'Why was that allowed?' — asked equally after correct and incorrect steps, so it stops signalling errors.
• After an answer: 'How confident are you, and what would make you more sure?'
• After a correct solution: 'Could you have done it another way? Which way is better here, and why?'
• To generalise: 'What would change in this problem if the numbers were different? What would break your method?'

Checking understanding without asking about understanding

'Do you understand?' fails because it asks the student to assess their own comprehension — the exact skill weak students lack, and the exact question social pressure answers for them. Replace self-report with performance. Instead of asking whether they understood the method, ask them to do something only understanding makes possible.

1. Explain it back: 'Teach me that step as if I had never seen it.' Gaps surface within one sentence.
2. Transfer it: give a near-variant with one feature changed and watch whether the method survives the disguise.
3. Predict: 'Before you calculate — roughly what should the answer be, and why?' Prediction exposes whether the structure is understood or just the procedure.
4. Find the error: show a worked solution containing one deliberate mistake and ask them to locate it. Spotting errors requires the very understanding that yes-saying conceals.

Every question a tutor asks is either doing the student's thinking or revealing it. The craft is in knowing, moment by moment, which one is happening.

None of this means abandoning direct explanation — novices need telling as well as asking, and Socratic questioning on content a student has never met is a frustration engine. The sequencing matters: explain and model when material is new, then shift steadily toward questions as competence grows. In that sense questioning follows the same fading logic as every other form of support. Early on you supply the thinking; by the end, your questions should be doing little more than confirming the student no longer needs them.