The Tutor Handbook Vol No.0097 | The Follow-Up Question

The Tutor Handbook · Volume 0097 · Series ID THB-0097

The Tutor Handbook: Complete series index.

The learner gives an answer that is not quite wrong.

That is often harder for a tutor than a clean error. Alicia chooses the correct algebraic method but cannot explain why it fits. Beatrice gives a defensible inference but points to evidence that is too broad. Ciara states the right scientific conclusion but skips the mechanism. Denise begins a proof in the right direction, then stops. Emily has a workable study plan but cannot say what she will do if the first task takes twice as long as expected.

The tutor now has several options. Confirm the good part. Correct the weak part. Explain the missing reasoning. Ask a broad “why?” Ask a narrower question. Give a hint. Move to another learner. Or say nothing and wait.

The next tutor move matters because a partly correct answer contains information and uncertainty at the same time. A poor follow-up question can reveal the answer, turn the exchange into guessing, or make the learner imitate the tutor’s reasoning. A well-chosen follow-up can expose the structure that is already present, locate the missing link and leave the intellectual work with the learner.

The Follow-Up Question is the tutor’s next question after a learner response, chosen to reveal, extend or test the learner’s reasoning without smuggling in the route the tutor wants to see.

This is not a general article about questioning. It does not replace The Diagnostic Probe, which owns the design of a task intended to separate competing explanations. It does not replace broader pages on teaching or feedback. This volume owns a moment-to-moment tutoring decision: the learner has already responded; what question should come next?

Quick Answer

A good follow-up question responds to the learner’s actual answer, not to the tutor’s prepared script. It should have a clear job: ask for evidence, justification, representation, comparison, prediction, checking, or clarification. The tutor should avoid questions whose wording contains the missing step, questions so broad that the learner cannot tell what to inspect, and rapid sequences that turn the interaction into a funnel toward the tutor’s answer.

Current educational guidance supports the underlying principles. Stanford’s National Student Support Accelerator recommends open-ended questions that make student thinking visible and give students substantial airtime. The OECD’s 2025 synthesis of classroom interaction highlights contingent follow-up questions that ask students to explain reasons and how they worked something out, and it notes evidence linking responsive moves such as revoicing and probing questions to deeper student contributions. These findings come mostly from classroom research rather than three-student Singapore tuition, so this volume treats them as a research-informed foundation, not proof that one scripted tutoring protocol will work everywhere.

1. The follow-up question begins after evidence arrives

The first question launches a task. The follow-up question must respond to what actually came back.

That distinction is important. Tutors often prepare a sequence: “What is the answer? Why? Which formula? What happens next?” A prepared sequence can support lesson planning, but if the second question is asked regardless of the first answer, it is not truly contingent. It may miss the learner’s actual reasoning.

Suppose a learner says, “I divided by three because there are three groups.” The tutor planned to ask, “Which operation do we use?” That question is now redundant. A better follow-up might be, “What tells you the groups are equal?” or “If the groups were not equal, would division still represent the situation?” The response has changed the next useful question.

Contingency is therefore the defining feature. The follow-up should be caused by the learner’s answer.

2. A partly correct answer has at least three layers

Before asking another question, separate what the learner has already demonstrated.

First is the answer layer: is the final response correct, incorrect or incomplete?

Second is the reasoning layer: what relationship, principle or evidence appears to support the response?

Third is the support layer: how much of the route came from the learner, the task, previous prompting, a model, a peer or the tutor?

A correct answer with weak reasoning calls for a different follow-up from an incorrect answer built on a strong principle and one arithmetic slip. A correct answer produced after a strong hint calls for a different follow-up from the same answer produced independently.

The tutor should therefore listen for what is already owned before deciding what to ask next.

3. Give the follow-up one job

A follow-up question becomes weak when it tries to do too much. “Why did you do that, is there another way, and how do you know it works?” contains several possible tasks. The learner may answer the easiest part and leave the tutor unsure what was learned.

Give the question one primary job. Useful jobs include:

Clarify: “When you say ‘it’, what exactly are you referring to?”
Justify: “What makes that step valid?”
Locate evidence: “Which words in the passage support that inference most directly?”
Compare: “What is different between this case and the previous one?”
Predict: “If this value doubled, what would you expect to change?”
Represent: “Can you show the same relationship with a diagram?”
Check: “What could you do to test whether this answer fits the question?”
Extend: “When would this method stop working?”

The question can be open while still being focused. Open-ended does not have to mean vague.

4. Do not put the missing answer inside the question

Tutors are experts at making hints feel like questions.

“Should you perhaps use the total before calculating the percentage?” is grammatically a question but instructionally close to a directive. “Does the author feel anxious because she keeps looking at the door?” already supplies the inference and the evidence. “What happens to the water vapour when it touches the cold surface?” may be appropriate during instruction, but if the target is for the learner to identify the relevant process independently, the wording can reveal too much.

The boundary depends on the learning job. A novice who has never learned condensation may need explicit instruction. A learner being tested for independent transfer needs less embedded help.

A useful self-check for tutors is: If I removed the question mark, how much of my sentence would still tell the learner what to do? If most of it remains a direction, the follow-up may be a disguised hint.

5. Avoid the guessing funnel

A guessing funnel looks like this:

“Is it multiplication?” “No.” “Division?” “No.” “Addition?” “Yes.”

The learner reaches the correct operation, but the exchange has sampled recognition among tutor-supplied options rather than method selection. The tutor may leave with an inflated impression of understanding.

Funnelling can happen more subtly. “What do we usually do when we see ‘of’?” “Which chapter is this from?” “Do you remember yesterday’s example?” Each cue narrows the route. Sometimes that is exactly the support a learner needs during guided practice. The problem arises when the tutor later treats the resulting answer as evidence of independent choice.

The follow-up question should therefore be matched to the support claim. If the tutor is teaching, a funnel may be a scaffold. If the tutor is checking what the learner can select alone, it can destroy the evidence.

6. Avoid the empty “why?”

“Why?” is powerful when the learner knows what claim is being examined. It can also be too broad.

A learner says, “The answer is 24.” The tutor asks, “Why?” The learner may not know whether to justify the operation, the arithmetic, the interpretation of the question or the reason the final unit is correct.

A more useful follow-up identifies the decision without supplying the answer: “Why did you multiply rather than add?” “What in the question tells you 24 represents the total?” “Which relationship are you using here?”

The OECD’s 2025 classroom-interaction chapter explicitly cautions that follow-up questions can become so broad that students are overwhelmed or sent in an unproductive direction. The point is not to avoid openness. It is to give the learner a productive object for reasoning.

7. Ask about the learner’s route, not merely the tutor’s preferred route

A tutor may see an elegant method and ask questions that steer the learner toward it even when the learner’s existing method is valid. This can create a false diagnosis: the learner appears unable to answer the follow-up because the tutor has switched the problem from “solve this” to “discover my preferred method.”

Before probing, ask whether the learner’s route is valid. If it is valid but inefficient, efficiency can become a separate teaching goal. If the route is invalid, identify where validity breaks. If several methods are acceptable, a comparative follow-up can deepen understanding without rewriting ownership.

“What would your method look like if the numbers were much larger?” can open an efficiency discussion. “Which step depends on these numbers being whole?” can reveal boundary conditions. The learner’s thinking remains the object.

8. Listen long enough to know what to ask

The quality of the follow-up depends on listening. Tutors often begin composing the next question while the learner is still speaking. The result is a question connected to the topic but not to the reasoning.

Stanford’s tutoring guidance emphasises giving students more airtime and listening for misconceptions and understanding. In small groups it explicitly encourages comfort with silence. These are facilitation principles rather than a rule that longer silence always causes more learning. Their practical value here is simple: the tutor needs enough of the learner’s thought to choose a contingent move.

A short pause after the answer can also prevent reflexive correction. The tutor can ask internally: What did I actually hear? What is still missing? What question would distinguish between the two most plausible interpretations?

9. The “tell me more” question has a limit

“Tell me more” can be excellent when the learner has a coherent idea that has not yet been fully expressed. It preserves ownership and invites elaboration. It becomes weak when the learner is genuinely lost.

If a student says, “I don’t know, I just guessed,” repeatedly asking “tell me more” can become performative. The tutor needs to change the level of support. A narrower question, representation, example, explanation or prerequisite check may be warranted.

The follow-up is not a ritual. It is a diagnostic move. The tutor should be willing to stop questioning and teach when the evidence says a knowledge gap, rather than a communication gap, is blocking progress.

10. Composite Mathematics case: correct method, weak justification

This case is fictional.

Alicia is solving a simultaneous-equations problem. She substitutes one equation into the other and obtains the correct answer. The tutor asks, “Why did you substitute?” Alicia says, “Because that’s what we do.”

There are several possible follow-ups. “Because one equation has y alone, right?” would feed the justification. “Can you explain more?” may be too broad. The tutor chooses: “What is true about this equation that makes substitution convenient here?”

Alicia points out that one variable is already isolated. The tutor then asks a changed-condition follow-up: “If neither variable were isolated, would substitution still be possible?” Now the conversation separates possibility from convenience.

The first follow-up revealed the selection criterion. The second tested whether Alicia understood the method’s boundary. Neither required the tutor to announce the missing reason.

11. Composite English case: inference is plausible, evidence is broad

Beatrice reads a passage and says a character is uncomfortable. Her evidence is “the whole paragraph where she is at the party.” The inference is plausible, but the evidence is not discriminating.

The tutor does not say, “Look at the line where she keeps checking the door.” Instead: “If you could keep only five words from that paragraph to defend ‘uncomfortable’, which five would you keep?”

This follow-up changes the task from finding relevant material to ranking evidence by directness. Beatrice chooses a phrase about the character standing near the exit. The tutor then asks, “What does that behaviour suggest, and what else could it mean?” The second follow-up protects against overclaiming.

The exchange is not a generic comprehension lesson. It is a sequence in which each question is caused by the quality of the preceding answer.

12. Composite Science case: right conclusion, missing mechanism

Ciara says that the water level decreased because of evaporation. The conclusion is correct. Her explanation is one word.

A weak follow-up is “What happens to the water particles when they gain heat?” because it nearly supplies the desired mechanism structure. A stronger first probe is: “What changed between the water in the container and the water that is no longer there?”

If Ciara says, “It became water vapour and went into the air,” the tutor can ask, “What would we need to observe to distinguish evaporation from water leaking out?” Now the learner connects mechanism to alternative explanation.

If Ciara cannot say what evaporation involves, the tutor should stop probing and teach. A follow-up question is not a substitute for missing knowledge.

13. Composite studying case: the plan is sensible but brittle

Emily creates a study plan for three evenings. The priorities are reasonable. The tutor notices there is no contingency for unfinished schoolwork.

Rather than say “add a buffer,” the tutor asks: “Which part of this plan breaks first if Tuesday’s homework takes an extra hour?”

Emily identifies the retrieval block as the likely casualty. The tutor follows: “Is that the block you most want to sacrifice, or simply the easiest one to push out?” The question reveals that the plan is prioritising urgency over retention by accident.

Emily then redesigns the plan. The follow-up did not merely improve a timetable; it made the learner inspect the rule beneath her own allocation.

14. A follow-up can test certainty without shaming uncertainty

Learners often answer with confidence levels that do not match the strength of their evidence. A tutor can probe certainty without turning confidence into a moral issue.

“What would make you change your mind?” is useful after an interpretation or prediction. “Which part are you most certain about?” can separate a strong principle from a weak final step. “If this answer were wrong, where would you look first?” can turn checking into a reasoning task.

These questions work best in a culture where changing one’s mind is normal. If every changed answer is treated as failure, learners will defend weak claims. The follow-up question is therefore embedded in relationship and classroom norms.

15. Use follow-ups to distinguish knowledge from expression

Some learners know more than they can initially say. Others speak fluently around a concept they do not control. A follow-up can help distinguish these states.

If a learner’s explanation is vague, ask for another representation: “Can you draw the relationship?” or “Show me with numbers.” If the diagram is precise, the issue may be verbal expression rather than conceptual understanding. If every representation remains vague, the knowledge itself may be unstable.

This is not a universal diagnostic test. Representation skill is itself a capability, and a learner may struggle to draw despite understanding. The point is to change the evidence condition carefully rather than infer too much from one mode of expression.

16. Use follow-ups to expose hidden support

A learner gives a polished answer from homework. The tutor asks, “How did you decide this was the important evidence?” The learner says, “My mother pointed to that line.”

The follow-up has uncovered support provenance. The answer may still be educationally useful, but it is not independent evidence of evidence selection. The tutor can now ask the learner to try a fresh passage.

Questions about process—“What did you look at first?”, “Where did you get stuck?”, “What help did you use?”—are not accusations. They help interpret the performance conditions. This connects to The Support Provenance Check.

17. The follow-up question can be closed

Open questions are valuable, but closed questions are not inferior by definition. The immediate learning goal matters.

“Did you use the original price or the discounted price as the percentage base?” can efficiently locate a specific branch after the tutor already knows the learner understands both concepts. “Does this evidence come before or after the event?” can check chronology. “Is this quantity constant?” can focus attention on a critical condition.

The danger is using closed questions to create a trail of yes/no cues that the learner can follow without constructing the reasoning. A closed follow-up is strongest when it resolves a specific uncertainty, not when it replaces the learner’s decision-making.

OECD’s synthesis makes a similar point: open and closed questions serve different purposes, and quality depends on the cognitive demand and the way responses are used.

18. The follow-up question can be a request for counterevidence

One powerful move is to ask the learner to challenge their own answer.

“What evidence in the passage would make your inference less likely?” “Can you find a case where this formula would not apply?” “What result would contradict your explanation?” These questions move beyond justification into boundary testing.

They are especially useful with strong learners who can produce polished explanations too quickly. The tutor is not adding harder content merely for difficulty. The learner is learning to keep a claim correctable.

But counterevidence questions require enough knowledge to reason about alternatives. With a novice, they can create confusion. The tutor should use them when the learner has a stable core model.

19. The follow-up question can return to an earlier step

The final answer is wrong, but the tutor does not need to correct the final line. Ask the learner to point to the last step they are certain about. Then inspect the next transition.

This creates a backwards diagnostic path without revealing the fix. It is particularly useful in multi-step Mathematics, extended writing and scientific explanations.

The learner may discover that uncertainty began much earlier than the visible error. The tutor gains a cleaner view of the first weak link. If the learner cannot locate any point of certainty, the task may be too complex for local questioning and require a simpler representative example.

20. In a three-student group, the follow-up can move sideways

A tutor does not always need to keep the question with the first learner. After Alicia explains a method, the tutor can ask Beatrice, “What part of Alicia’s reasoning would you keep if the numbers changed?” or ask Ciara, “Can you give a case where that method would not fit?”

This uses peers to extend the reasoning without asking one learner to carry the entire dialogue. It also makes listening consequential: students need to engage with each other’s ideas, not only wait for their turn.

However, Volume 0094’s peer-answer leakage boundary still applies. If Beatrice has not yet produced independent evidence on the same target, hearing Alicia’s full route may contaminate the next check. Group dialogue and independent evidence have to be sequenced deliberately.

21. Do not interrogate the learner into exhaustion

Tutors can overvalue questioning. A learner who has answered six successive probes may feel that every response merely unlocks another test. The dialogue becomes adversarial even when the tutor’s intention is intellectual depth.

Use follow-ups until they resolve the current instructional uncertainty or create the desired reasoning opportunity. Then teach, practise, summarise or move on. Not every correct answer needs a deeper question. Sometimes a concise confirmation is the efficient response.

A practical rule is: questions should earn their cognitive cost. Ask because the answer will change what the tutor or learner does next.

22. Do not ask a question whose answer you will ignore

A tutor asks, “Which part do you find confusing?” The learner says the vocabulary. The tutor proceeds with a method explanation because that was the prepared plan. The question has become theatre.

If learner voice is invited, it must have some chance of changing the route. That does not mean the learner controls every decision. It means the tutor treats the response as evidence.

When the tutor disagrees, explain why. “You said the vocabulary is the problem. Let’s check that first. You understood all the relationship words, but the representation still broke, so I think the next repair is the diagram rather than more vocabulary.”

The learner sees that their answer was heard and tested.

23. Build a small library of question jobs, not scripts

Scripts can help novice tutors, especially during training. But scripts become brittle if they are followed independent of student response.

A better professional-learning approach is to rehearse question jobs. Given a student answer, can the tutor choose whether they need clarification, justification, evidence, comparison, prediction, representation, checking or extension? Then can they phrase one question that serves that job without embedding the answer?

Stanford’s practice-based tutor-training framework recommends rehearsal and feedback rather than relying only on lectures or workshops. That supports training tutors on live decision-making. The exact taxonomy in this article is an eduKate tutoring proposal, not a Stanford standard.

24. Research foundation and limits

The OECD’s 2025 Unlocking High-Quality Teaching chapter on classroom interaction synthesises research on questioning, responding, collaboration and discussion. It highlights contingent follow-up questions that ask learners to give reasons or explain how they worked something out. The chapter also cites a study by Bishop involving a convenience sample of 13 teachers and 250 students in which highly responsive teacher moves, including revoicing and follow-up questions, were positively associated with learning of the focal Mathematics topic. That finding is informative but should not be treated as universal causal proof for all tutoring settings.

Stanford’s National Student Support Accelerator recommends open-ended questions in tutoring to make student thinking visible, increase student airtime and support productive struggle. Its one-to-one facilitation checklist gives examples that ask learners to explain strategies, identify difficult parts and inspect possible mistakes. These are research-informed implementation resources for tutoring, not randomised evidence that one wording is always superior.

AERO’s Monitor Progress resources emphasise frequent checks for understanding, guiding attempts and responding constructively to mistakes. Together, these sources support responsive questioning as part of formative instruction. They do not validate a fixed sequence of follow-up questions, and tutors should remain sensitive to age, subject, access needs, language, learner confidence and the difference between instruction and assessment.

25. The Follow-Up Question card

What has the learner already demonstrated?
What is still uncertain?
What is the one job of my next question?
Does the wording reveal the missing step?
Am I probing the learner’s route or steering toward my preferred route?
Is an open question appropriate, or would a focused closed question resolve the uncertainty more cleanly?
Is the learner missing knowledge that requires teaching rather than more probing?
Has the learner had enough thinking time?
Will the answer change what I do next?
Is peer involvement useful, or would it contaminate independent evidence?
Have I asked enough questions already?
What fresh task will show whether the reasoning survives beyond this dialogue?

If the next question has no clear informational or learning value, do not ask it merely because good tutors are supposed to ask questions.

26. The parent’s view

Parents listening to tuition may hear many questions and assume the tutor is withholding help. Or they may hear fast explanations and assume the tutor is efficient. Neither surface pattern is enough.

Useful questions for parents are: Does the tutor let the child explain? Do questions respond to what the child actually says? Does the tutor know when to stop questioning and teach? Does the child become better at explaining decisions, not merely better at guessing what the tutor wants?

A strong tutoring dialogue should gradually make the learner’s reasoning more visible and more self-directed. It should not create dependency on the tutor’s next clue.

27. The learner’s view

The learner should understand why follow-up questions exist. “I’m not asking because your first answer was bad. I’m asking because I want to see which part you already control and which part needs work.”

That explanation changes the social meaning of questioning. The tutor is not cross-examining the learner. The learner is helping inspect their own reasoning.

Over time, the best follow-up questions can migrate inward. The learner begins to ask themselves: What evidence supports this? Why does this method fit? What would make this wrong? Can I show it another way? That is where tutor questioning starts becoming learner self-regulation.

A follow-up question has done its deepest work when the tutor no longer needs to ask it.

Evidence and connected reading

Final compression

The learner answers. The tutor should not merely continue the script.

Listen for what is already owned. Decide what remains uncertain. Give the next question one job. Do not hide the answer inside the wording. Avoid guessing funnels. Avoid empty “why?” questions. Probe the learner’s route, not your preferred route. Change from broad to narrow support when the learner is genuinely stuck. Use questions to clarify, justify, locate evidence, compare, predict, represent, check or extend. Stop questioning when teaching is the better move.

Then ask whether the reasoning survives a fresh task without the dialogue that produced it.

That is the Follow-Up Question.

That is Tutor Handbook Volume 0097.

Small Group Tutorials

The Tutor Handbook Vol No.0097 | The Follow-Up Question — How a Tutor Responds to a Partly Correct Answer Without Feeding the Learner the Route