eduKateSG Learning Node Series · 0081
How Teacher Inquiry Works | Turn Classroom Questions Into Evidence, Action and Professional Learning
A teacher can finish five lessons with a strong feeling that something is wrong and still not know what the problem is.
The class was attentive. The examples were clear. Homework completion looked respectable. Then the assessment arrived and a pattern appeared: many students could imitate the method but could not recognise when to use it. The easy reaction is to teach the topic again. Teacher inquiry asks a harder question first: what, exactly, failed to travel from instruction into independent performance?
Teacher inquiry is the disciplined habit of turning a problem from practice into a question, gathering evidence close to the learning, forming and testing an explanation, changing something deliberately, and returning to see what happened. It is not a miniature university thesis performed after school. It is professional learning built around the classroom itself.
The classroom produces signals every day. Inquiry begins when the teacher stops treating those signals as impressions and starts turning them into testable professional questions.
The 50-Second Read
- Teacher inquiry is a systematic investigation of a real problem of teaching or learning carried out by practitioners close to the work.
- It begins with a problem worth understanding, not with a fashionable intervention looking for somewhere to be used.
- Useful inquiry questions are narrow enough to inspect and important enough to change practice.
- Evidence may include student work, assessment responses, observations, short interviews, learner explanations, classroom artefacts and carefully chosen quantitative measures.
- The aim is not to prove the teacher right. The aim is to improve the teacher’s model of what is happening.
- Inquiry improves when teachers distinguish observation from explanation: “students omitted evidence” is not the same claim as “students were careless.”
- Small changes are often easier to learn from than large bundles of simultaneous changes.
- Collaboration, external expertise, leadership support and sufficient time strengthen teacher inquiry.
- A 2020 integrative review reported that effective teacher inquiries associated with positive student-achievement shifts tended to include a student-achievement focus, time, teacher engagement, collaboration, external expertise and leadership support.
- A 2019 review of action research described systematic practitioner inquiry as a route to intentional change in pedagogical content knowledge, disciplinary inquiry and practice, while noting logistical and dissemination barriers.
- Teacher inquiry does not replace educational research. It gives teachers a disciplined local learning mechanism that can work with research rather than merely receive it.
- The strongest inquiry ends with a better next question, not a ceremonial conclusion.
Canonical Owner Boundary
This page owns the teacher inquiry mechanism: a practitioner-led cycle in which a classroom problem becomes a question, evidence, interpretation, deliberate change and return. How Education Works | Teacher Professional Learning remains the umbrella owner for career-long teacher development. How Education Works | Educational Research owns the wider production and interpretation of educational knowledge. How Lesson Study Works owns collaborative inquiry organised around a jointly designed research lesson. How Professional Learning Communities Work owns the recurring community architecture for collective professional learning. Teacher inquiry may occur inside any of these systems, but its distinctive unit is the practitioner question and its evidence-return loop.
Worked inquiry case: question → evidence → teaching change → later observation
This is an illustrative teaching case, not a claim about a real class. It shows the decision structure Teacher Inquiry should preserve.
Question
A teacher notices that students can calculate percentage change in labelled exercises but often choose the wrong base when the same relationship appears inside unfamiliar wording. The inquiry question becomes: is the difficulty mainly weak percentage knowledge, or failure to identify the reference quantity when the surface changes?
Evidence
The teacher uses a short mixed task with three item types: a familiar percentage exercise, an unfamiliar context requiring the same percentage relationship, and a near non-example where the tempting denominator is wrong. The teacher records not only answers but the quantity each learner chose as the base and a one-sentence reason.
Suppose the illustrative pattern is: most learners calculate accurately once the base is supplied, but several choose the wrong base when they must identify it themselves. That evidence supports a narrower explanation than “they do not know percentages.”
Alternative explanation
A competing explanation is language load: perhaps the learners understand the relationship but lose it when the sentence structure is unfamiliar. The teacher therefore checks one visually presented version of the same relationship and one language-heavy version. If errors cluster only in the language-heavy item, the original explanation should be revised rather than forcing a mathematics repair onto a comprehension problem.
Teaching change
The teacher changes one thing deliberately: instead of more identical percentage questions, students compare examples where the numerical operations look similar but the reference quantity changes. They must mark the base before calculating and explain why it is the correct denominator. The subject-specific mathematics remains with the specialist; this page owns the inquiry logic behind deciding what to change.
Later observation
After a delay, the teacher returns with a mixed changed-task set in a different context and again records base choice plus explanation. If the specific wrong-base error falls while other errors remain stable, that is evidence that the targeted teaching change helped this defined problem. If performance improves only on familiar wording, the explanation remains incomplete and the next inquiry cycle should test language support or transfer rather than declaring the issue solved.
The conclusion is deliberately bounded: one classroom pattern can justify the next teaching move without becoming universal proof. Learner-specific implementation and ongoing observation belong with the Sengkang runtime; subject claims should be checked against the relevant specialist owner before being generalised.
1. Inquiry Starts With Friction
The most useful inquiry questions often begin with a small professional irritation. Students can define a term but misuse it in explanation. They solve the practice set but freeze when the representation changes. They revise carefully yet repeat the same error two weeks later. Discussion is lively, but written reasoning remains thin.
Friction matters because it points to a mismatch between what the teacher expected the learning system to produce and what students actually produced. Inquiry makes that mismatch inspectable.
2. A Problem Is Not Yet an Inquiry Question
“My class is weak at comprehension” is too broad. It bundles vocabulary, inference, attention, background knowledge, question interpretation, evidence selection and written expression into one label.
A stronger question might be: “When students answer inference questions correctly in discussion but lose marks in writing, where does the reasoning chain break?” Now evidence can be gathered around a specific transition.
3. Ask a Question That Can Change a Decision
An inquiry question earns its cost when different answers would lead to different teaching. If every possible finding leads to the same lesson plan, the question may be academically interesting but operationally weak.
“Do students like group work?” may tell us something. “Which part of a group problem-solving routine causes weaker students to stop producing independent reasoning?” is more likely to change design.
4. Separate Observation From Explanation
Teachers are experienced pattern recognisers, which is powerful and dangerous. A teacher sees eight incomplete answers and immediately thinks: careless. But “incomplete answer” is an observation. “Careless” is an explanation.
Other explanations may fit: students did not understand the command word, could not retrieve the second step, ran out of time, misread the diagram, or believed one example was sufficient. Inquiry keeps the claim open long enough to discriminate among possibilities.
5. Build a Small Evidence Set Before Building a Big Story
Teachers already possess rich evidence: marked scripts, oral responses, exit tickets, notebooks, recordings, question-level scores, student annotations and the sequence of prompts used in a lesson.
The challenge is not collecting everything. It is choosing evidence that can distinguish explanations. Three carefully selected student responses may reveal more about a misconception than a spreadsheet containing 1,000 undifferentiated scores.
6. Sample for Contrast, Not Only Convenience
If the question concerns why some learners transfer a method and others do not, compare cases. Choose one student who transferred successfully, one who almost did and one who failed despite apparently similar practice.
Contrast makes hidden differences visible. Perhaps the successful student represented the problem before calculating. Perhaps the weaker student memorised a surface cue. The inquiry now has leverage.
7. Student Explanations Can Reveal the Invisible Step
A written answer shows what was produced. A short learner explanation can reveal why. Ask a student to reconstruct the decision: “What made you choose this formula?” “What did you think this question was asking?” “Where did you first become unsure?”
The purpose is not interrogation. It is cognitive archaeology. The teacher is trying to reconstruct the path that produced the visible answer.
8. Student Voice Is Evidence, Not a Verdict
Learners can report that an explanation felt clear while retaining a misconception. They can dislike a demanding routine that improves learning or enjoy an activity that produces little durable knowledge.
Student voice matters because learners experience parts of instruction teachers cannot directly see. But it should be triangulated with performance, observation and task evidence rather than treated as a customer-satisfaction score.
A 2024 systematic review of 19 K–12 studies found that student voice can stimulate teacher reflection and practice change, particularly when supportive leadership and collaborative structures help teachers act on what learners report. That evidence supports using student voice as a professional-learning input, not as a standalone causal measure.
9. The Teacher Needs a Working Theory
Evidence becomes useful when it changes a model. The teacher might conclude: students know the procedure but rely on a surface cue that disappears in novel questions. Or: students understand the concept orally but cannot compress it into examination language. Or: retrieval is stable at one week but collapses after three.
A working theory is provisional. It should be strong enough to guide an intervention and weak enough to be changed by new evidence.
10. Competing Explanations Protect the Inquiry
Before changing instruction, list at least two plausible explanations. If students fail a new algebra problem, is the bottleneck algebraic manipulation, representation, reading load, or method selection?
Competing explanations prevent the teacher from collecting only evidence that supports the first story. Inquiry should be capable of surprising its owner.
11. External Research Expands the Hypothesis Space
Teacher inquiry becomes stronger when local experience meets wider knowledge. Research may introduce mechanisms the teacher had not considered: retrieval difficulty, cognitive load, feedback timing, prior knowledge, representational competence, motivation, language demands or implementation conditions.
Research should not dictate the classroom mechanically. It should widen and constrain the explanations worth testing.
12. Inquiry Is Not “Try Something and See”
Teachers experiment informally every day. Inquiry adds structure. It states what will change, why it might work, which evidence should respond and when the teacher will look again.
If we make students explain why two superficially different questions share the same structure before independent practice, then method selection should improve on unfamiliar problems.
Now the change has a mechanism and a predicted signal.
13. Change One Important Thing at a Time When Possible
If a teacher changes the textbook, seating plan, feedback method, homework quantity and explanation style simultaneously, improvement may occur but the inquiry learns little about why.
Real classrooms cannot become laboratories with every variable controlled. Still, disciplined restraint helps. Change enough to matter, but not so much that interpretation becomes impossible.
14. Define the Evidence Before the Intervention
Teachers are vulnerable to seeing success after investing effort. Before implementation, decide what would count as improvement.
- More students select relevant evidence without prompting.
- Errors shift from conceptual to procedural.
- Performance survives a delayed retest.
- Students solve a structurally similar problem with different surface features.
- The same quality appears with fewer teacher prompts.
Pre-specifying the signal makes the later interpretation more honest.
15. Immediate Improvement Is Not the Only Improvement
A new scaffold may improve performance during the lesson and reduce independence later. A slower explanation may reduce same-day completion while strengthening understanding. A retrieval routine may initially make students feel less fluent while improving durable access.
Inquiry should measure the capability the teacher actually cares about, at the time horizon that matters.
16. Look for Transfer
If students improve only on the exact task used during the intervention, the teacher may have trained familiarity rather than the underlying capability.
Change the numbers, context, representation, wording or source text. The closer the new task is to the intended future use, the stronger the evidence that learning travelled.
17. Look for Independence
Support can make performance look better while capability remains unchanged. Record how much help students needed. If the answer quality rises but teacher prompting remains constant, the intervention may not yet have transferred control to the learner.
Strong inquiry asks not only “Can they do it?” but “Can they initiate and sustain it themselves?”
18. Negative Results Are Valuable Professional Knowledge
A thoughtful intervention can fail. Perhaps the hypothesised mechanism was wrong. Perhaps implementation was weak. Perhaps the effect appeared only for learners with sufficient prior knowledge. Perhaps the measure could not detect the change.
Failure is useful when it narrows the model. Hiding negative results forces the profession to rediscover the same dead ends.
19. Implementation Evidence Belongs Inside the Inquiry
If the intervention did not work, first ask whether it actually happened as intended. How often was the routine used? Were students trained in it? Did time pressure remove the key step? Did teachers adapt away the active ingredient?
Outcome data without implementation data can blame an idea for a delivery failure—or celebrate an idea when something else caused the improvement.
20. Inquiry Should Be Proportionate
Not every classroom decision needs a research project. A teacher deciding whether to move one example earlier in a lesson can often use professional judgement. Inquiry is most valuable when uncertainty is consequential, recurring or resistant to ordinary adjustment.
The method should match the stakes. Do not build a twelve-week data system to answer a question that can be resolved by examining six scripts tomorrow.
21. Ethics Still Apply to Practitioner Inquiry
Teachers hold authority over students and access to sensitive information. Inquiry should protect confidentiality, avoid unnecessary risk, minimise disruption and follow school or institutional requirements for consent and data use.
When work moves from ordinary professional reflection toward formal research, publication or data sharing, ethical review requirements may change. Practitioner proximity is not exemption from responsibility.
22. Collaboration Improves Error Detection
A colleague can ask the question the inquiry owner stopped seeing: “How do you know they misunderstood rather than forgot?” “Could the new rubric be creating the apparent gain?” “Why did you sample only your strongest class?”
Collaboration is valuable not because agreement feels supportive but because another mind can expose assumptions.
23. External Expertise Prevents Local Circularity
The 2020 integrative review of teacher inquiry identified external expertise among the conditions associated with stronger inquiries. This makes sense. A school can become extremely sophisticated at discussing its own habits without encountering an alternative model.
Subject specialists, researchers, instructional coaches, experienced practitioners and high-quality literature can challenge the local explanation and introduce methods the team does not already possess.
24. Leadership Support Is Infrastructure, Not Decoration
Inquiry requires time to inspect evidence, read, collaborate, test and return. If teachers must conduct it entirely through unpaid extra labour, the school has not built an inquiry culture; it has built an enthusiasm tax.
Leadership can protect time, access to data, professional trust, expertise and permission to report uncomfortable findings. It can also destroy inquiry by turning every project into an accountability display.
25. Inquiry Dies When the Answer Is Politically Predetermined
If a school has already decided that a new programme must be successful, teachers are not conducting inquiry. They are collecting testimonials.
Real inquiry requires the possibility that the favoured explanation weakens, the intervention needs adaptation or the initiative should stop.
26. Cross-Domain Comparison: Clinical Case Review
In high-quality clinical work, a practitioner does not treat one symptom as a complete diagnosis. History, observations, tests and response to intervention update a working model. The model is provisional and revised as evidence changes.
Teacher inquiry can use the same epistemic discipline without pretending education is medicine. A learner response is evidence about a learning state, not the state itself. Multiple observations improve the model. An intervention can also become information when its success or failure changes the diagnosis.
27. Cross-Domain Comparison: Software Debugging
A software team facing an intermittent failure does not normally rewrite the entire system first. It reproduces the problem, inspects logs, isolates conditions, forms hypotheses, changes one component and watches what happens.
Teacher inquiry benefits from the same anti-chaos principle. Find the reproducible learning failure. Inspect the trace. Change the smallest meaningful mechanism. Retest under a condition that can disconfirm the hypothesis.
28. Cross-Domain Comparison: Engineering Failure Analysis
When a structure fails, engineers distinguish the visible break from the causal chain. Material weakness, load, geometry, fatigue, maintenance and operating conditions may interact.
Educational failure also has chains. A poor essay may arise from weak knowledge, planning, sentence control, task interpretation or time allocation. Inquiry asks which link is limiting performance rather than repairing every visible defect equally.
29. Example: Mathematics Method Selection
A teacher notices that students solve simultaneous equations reliably when the exercise heading names the method but struggle in mixed practice. The first hypothesis is forgetting. Inspection shows something else: many students cannot classify the structure when the question does not announce itself.
The teacher introduces short comparison prompts: “What features make elimination efficient here?” “How is this pair different?” Students must classify before solving. The later measure is not the same worksheet. It is an unfamiliar mixed set with no headings.
If method selection improves, the inquiry has learned something more precise than “extra practice works.” It has identified representation and classification as part of the missing capability.
30. Example: English Evidence and Explanation
A Secondary English teacher sees students quote accurately but make unsupported claims. The teacher initially plans more comprehension questions. Before doing so, she compares strong and weak responses and asks students to explain how they chose evidence.
The weak responses often contain relevant quotations. The failure lies in explaining the inferential bridge from quotation to claim. The intervention therefore changes: students compare two responses that use the same quotation but make different inferences, then identify which reasoning step is defensible.
The inquiry has moved the repair point from “find evidence” to “control what the evidence licenses.”
31. Example: Retrieval That Looks Strong Too Early
A class performs well on vocabulary quizzes conducted the day after teaching. A month later, the words disappear from writing. The teacher’s initial conclusion—students are not revising—may be true, but the inquiry tests a design question instead.
One group receives the normal next-day quiz plus delayed retrieval at one and three weeks. Another receives only the usual schedule. The teacher tracks not merely recognition but accurate use in new sentences.
The inquiry changes the time scale of evidence. What looked like successful teaching at twenty-four hours may have been fragile learning.
Worked Inquiry Decision Trace: One Question, One Change, One Later Check
This is an illustrative classroom case, not a reported eduKate dataset and not evidence of a universal effect. Its purpose is to show how one inquiry can stay bounded from question to evidence, teaching change, later observation and the next decision.
The question
A Mathematics teacher notices that students can solve simultaneous equations when the exercise heading names the method, yet some fail when elimination, substitution and unrelated algebra questions are mixed together. The inquiry question is narrower than “Why are they weak at algebra?”: When the method label disappears, is the main bottleneck procedural forgetting, representation, reading load or method selection?
The minimum evidence set
The teacher samples contrasting recent scripts: one learner who transferred successfully, one who almost did and one who chose the wrong method despite completing the procedure correctly in blocked practice. She also asks a short reconstruction question: “What feature of the equations made you choose this method?” The aim is not to collect more data than necessary. It is to gather evidence capable of separating plausible explanations.
- Explanation A — procedural forgetting: the elimination procedure is no longer stable.
- Explanation B — method selection: the procedure is available, but the learner relies on a surface cue such as the worksheet heading.
- Explanation C — representation or reading load: the learner can choose correctly when equations are already represented but struggles when the same relationship arrives in a different form.
If students still cannot perform elimination after the method is named, Explanation A remains plausible. If they can execute the procedure but cannot state why it fits, Explanation B gains weight. If errors cluster around dense wording rather than symbolic items, Explanation C remains live. The evidence narrows the next decision; it does not diagnose the learner.
The teaching change
The teacher changes one important thing. Before solving, students compare two superficially different simultaneous-equation problems and answer one classification prompt: “Which structural feature makes elimination efficient here, and what would make substitution preferable instead?” The comparison is modelled once, then prompts are reduced during independent work.
The mathematical method itself remains a subject claim that must be correct independently of the inquiry. Classroom implementation can be inspected locally, but the validity of the Mathematics belongs to the relevant subject standard and specialist expertise. Teacher inquiry does not create a new rule because one classroom pattern seems persuasive.
The later observation
The later check is defined before seeing the result. After a delay, students receive a mixed set with changed coefficients, different question order and no heading announcing the method. The teacher records three things separately: whether the learner classifies the problem appropriately, whether the chosen procedure is executed correctly and how much prompting remains necessary.
Suppose the illustrative follow-up shows that several learners now classify changed symbolic items correctly but still fail when the same relationship is embedded in a dense word problem. That pattern weakens a pure “they forgot elimination” account. It supports a narrower interpretation: the comparison prompt may have helped method classification under one representation, while translation or reading demand remains unresolved.
The alternative explanation stays alive
Even that improvement does not establish that the prompt caused the change. Students may have improved because of extra exposure, peer discussion, a simpler later set, stronger prerequisite algebra or ordinary learning across the unit. The inquiry therefore checks what else changed and whether the pattern repeats. A local observation is evidence for the next professional decision, not a universal causal claim.
The bounded decision
The teacher now has a more useful next action than “more practice.” Keep the short classification comparison because it is worth further testing; separate representation-heavy problems into a different question; reduce prompting on the symbolic set; and recheck on another delayed mixed task. If later evidence fails to repeat, revise the working theory instead of protecting it.
This is the scale at which teacher inquiry becomes operational: one consequential question, a small evidence set chosen to distinguish explanations, one meaningful teaching change, a later observation under changed conditions, an explicit alternative explanation and a next decision whose certainty does not exceed the evidence.
Local inquiry should change the next professional decision before it tries to make a universal claim.
32. The Research Base: Promising, Uneven and Context-Dependent
The evidence on teacher inquiry should be read carefully. A 2020 integrative review by Turner-Adams, Wilson and Jesson examined research connecting teacher inquiry with student achievement. It identified conditions associated with stronger inquiries: focus on student achievement, sufficient time, teacher engagement, collaboration, external expertise and leadership support. The review did not justify treating every inquiry project as automatically effective.
A 2019 review by Meghan Manfra described action research as a systematic and intentional approach to changing teaching practice, with evidence across English language arts, mathematics, science and social studies. It also highlighted barriers including logistical constraints, the marginal status sometimes given to practitioner research and difficulty disseminating findings.
A 2023 systematic review of teacher professional learning found that ongoing coaching, collaborative professional development and frequent, longer-duration learning appear promising for student outcomes, while also noting that only a small subset of the 125 reviewed studies used experimental or quasi-experimental designs to examine student learning. The evidence base therefore supports disciplined professional learning while warning against easy causal claims.
33. Failure Mode: The Question Is Too Large
“How can I improve motivation?” can consume a year and still produce fog. Narrow it: “What happens to completion and quality when students choose the order of three equally demanding practice tasks?”
A smaller question is not less ambitious. It creates a usable unit of learning.
34. Failure Mode: Data Collection Becomes the Work
Teachers build surveys, coding sheets, trackers and dashboards until the inquiry exists mainly to feed its own evidence system.
Use the minimum evidence that can change the decision. The classroom should remain a place for learning, not a data mine.
35. Failure Mode: Reflection Without Disconfirmation
Reflection can become a polished story explaining why the teacher’s preferred idea was successful. Inquiry needs stronger friction: what evidence would make us revise the story?
A professional who cannot name disconfirming evidence is not yet testing the explanation.
36. Failure Mode: The Project Ends at Presentation Day
Many inquiry programmes culminate in posters, slides or showcases. Useful—but the real endpoint should be changed professional memory. What should teachers do differently next term? Which uncertainty remains? What should be retested?
Presentation is dissemination. It is not closure.
37. A Practical Teacher Inquiry Cycle
- Name one recurring learning or teaching problem.
- Describe the observable pattern without diagnosing it yet.
- Write a narrow question whose answer could change practice.
- Collect a small baseline evidence set.
- Compare contrasting cases.
- Ask learners to reconstruct relevant decisions where useful.
- Generate at least two plausible explanations.
- Consult relevant research or external expertise.
- Choose a working theory.
- Design one meaningful change linked to that theory.
- Define the expected evidence before implementation.
- Record whether the change was actually delivered as intended.
- Collect outcome evidence at an appropriate time horizon.
- Test transfer or independence where those matter.
- Compare predicted and observed results.
- Keep, adapt or reject the hypothesis.
- Record what was learned and what remains uncertain.
- Share the finding with colleagues who face the same problem.
- Ask the next question from the new state of knowledge.
38. A Lightweight Inquiry Record
- Problem: What recurring learning difficulty matters?
- Observation: What can we actually see?
- Question: What are we trying to understand?
- Competing explanations: What else could produce the pattern?
- Evidence: Which small set can distinguish those explanations?
- Working theory: What currently seems most plausible?
- Change: What will we do differently?
- Prediction: What should change if the theory is right?
- Implementation check: Did the planned change really occur?
- Result: What changed, for whom and under which conditions?
- Decision: Keep, adapt, stop or investigate further?
- Next uncertainty: What do we still not know?
39. Missing-Node Scan: Questions a Mature Inquiry Culture Should Ask
- Can teachers identify observations separately from interpretations?
- Do inquiry questions point to decisions rather than broad themes?
- Can staff access student work and question-level evidence easily?
- Are learners sometimes asked to explain the route behind an answer?
- Does the school protect inquiry time rather than relying on volunteer overtime?
- Can teachers report that a favoured strategy failed without professional embarrassment?
- Does external research enter the process before conclusions harden?
- Can colleagues challenge the inquiry owner’s explanation?
- Are interventions small enough to learn from?
- Is implementation evidence collected as well as outcome evidence?
- Does the inquiry check delayed learning, transfer or independence where appropriate?
- Are negative results preserved?
- Can another teacher understand the reasoning behind a changed practice?
- Does useful local knowledge travel beyond the original classroom?
- Do completed inquiries generate better questions rather than administrative closure?
40. Authoritative Research Starting Points
Useful starting points include Turner-Adams, Wilson and Jesson’s 2020 integrative review of teacher inquiry and student achievement; Meghan McGlinn Manfra’s 2019 review of action research and systematic change in teaching practice; the 2024 systematic review of student voice in teacher professional learning; and a 2023 systematic review of teacher professional learning and student outcomes.
These sources converge on a practical message. Practitioner learning becomes more credible when it is sustained, evidence-connected, collaborative, supported and open to revision. They do not make every local inquiry scientifically generalisable. They make local professional reasoning more disciplined.
41. The Return Path
At the beginning, the teacher says, “They understand in class but cannot do it alone.”
After inquiry, that sentence becomes more precise. Students can execute the method when a familiar surface cue identifies it. They struggle when they must classify the problem first. Comparison prompts improve classification, but only when students explain the structural feature themselves. The gain survives a delayed mixed set for most students but not those with weak prerequisite algebra.
The teacher now possesses something better than a feeling and something more useful than a slogan. She has a sharper model, a tested move, a boundary condition and a next question.
Teacher inquiry works when practice becomes evidence, evidence improves the model, the model changes the next action, and the next action returns new evidence to the teacher.