VIEW THIS AS

Auto mode follows the Route Engine until you choose a viewpoint.

YOU ARE HERE

ROUTE CHECK

CONNECTED TO

WHAT NEXT

Use the canonical route for this room, or HELP if you are unsure.

Did the Teaching Change Work, or Did the Task Change? | A Teacher-Inquiry Guide to Alternative Explanations

Three secondary students review open books and notes together in a classroom.

eduKateSG · Research & Inquiry Hub · Did the Teaching Change Work, or Did the Task Change? · 2 October 2026

Teachers change something and then look for improvement. That is sensible. The danger is that classrooms keep changing even when the named intervention does not. The second task may be easier. Students may know the format. Attendance may shift. The teacher may give more cues. Another programme may begin. A low baseline may rebound naturally. The new method may still be good—but a score rise by itself cannot tell us which change produced it.

This guide develops SG-AUD-20260929-04 under eduKateSG’s Research & Inquiry Hub. It deepens How Teacher Inquiry Works without replacing that broader owner. The narrow job here is causal caution: when a teaching change and an outcome change occur together, what evidence lets a teacher say more than ‘they happened in the same sequence’?

The distinction between learning and immediate performance matters here too. Soderstrom and Bjork’s Learning Versus Performance review shows why same-day success can be a poor proxy for durable learning. Classroom research also reminds us that broad instructional labels can conceal major implementation differences. A 2022 longitudinal study of inquiry-based teaching, for example, found little positive association with science achievement in its context and discussed how frequency, quality, guidance and operational definition complicate interpretation. The lesson is not that one pedagogy is good or bad. It is that causal claims need a design that matches the question.

All worked numerical cases below are explicitly illustrative. They are not student results collected by eduKateSG. The purpose is to show how a teacher can reason about evidence without inventing a laboratory inside an ordinary classroom.

A teaching change and a score change are two events, not yet a causal chain

A teaching change and a score change are two events, not yet a causal chain is mainly about separating intervention from outcome before attributing one to the other. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is a class using a new retrieval routine and later scoring higher on a quiz. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is saying the routine caused the gain because it happened first. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that alternative explanations are listed before the conclusion. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Define what changed in the teaching

Define what changed in the teaching is mainly about making the intervention specific enough to inspect. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is replacing rereading with three short retrieval prompts at the start of each lesson. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is describing the intervention only as ‘more active learning’. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that another teacher could understand what was actually done. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Define what should change if the mechanism is right

Define what should change if the mechanism is right is mainly about stating a predicted signal before the intervention. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is expecting unaided recall after delay to improve if retrieval practice is the mechanism. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is waiting for any positive outcome and calling it success. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the predicted outcome matches the proposed mechanism. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

The outcome task must sample the intended capability

The outcome task must sample the intended capability is mainly about aligning measurement with the teaching goal. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is using fresh recall and application items after a retrieval intervention. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is measuring only class enjoyment after a memory-focused change. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the evidence addresses the claimed capability. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Task difficulty must be comparable

Task difficulty must be comparable is mainly about checking whether post-intervention items became easier. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is matching cognitive demand, content coverage and format across checks. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is comparing a difficult baseline paper with an easier follow-up. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the result is interpreted with task demand. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Task format can create false improvement

Task format can create false improvement is mainly about recognising gains from familiarity with question style. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is using a different but construct-equivalent format as a robustness check. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is repeating the exact worksheet pattern. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that format learning is separated from subject learning. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Support conditions must remain visible

Support conditions must remain visible is mainly about tracking hints, notes, peer help and teacher prompts. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is comparing two tasks completed under similar support levels. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is allowing much more prompting after the intervention. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that independence is part of the evidence. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Practice effects can mimic teaching effects

Practice effects can mimic teaching effects is mainly about recognising that repeated testing itself can improve performance. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is using fresh items and acknowledging retrieval from the baseline test. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is crediting every retest gain to the new instructional routine. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the design considers exposure from earlier measurement. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Regression to the mean matters after unusually low results

Regression to the mean matters after unusually low results is mainly about avoiding causal claims when intervention starts after an extreme score. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is a class receiving a new strategy after one unusually poor test. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is attributing the natural rebound entirely to the strategy. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that multiple baseline observations reduce the risk. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Normal maturation and ongoing instruction continue

Normal maturation and ongoing instruction continue is mainly about recognising that students learn from many sources over time. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is noting that the class also continued regular lessons during the intervention. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is pretending the new routine was the only learning input. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the interpretation acknowledges ordinary progression. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Concurrent changes create confounding

Concurrent changes create confounding is mainly about tracking other interventions introduced at the same time. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is a new textbook, seating plan and homework policy arriving alongside the teaching change. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is crediting the observed outcome to one preferred change. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the design isolates or documents co-changes. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Attendance changes can alter class results

Attendance changes can alter class results is mainly about checking who was present in each measurement. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is a follow-up test missing several frequently absent low scorers. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is comparing class means without participant consistency. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the analysis notes matched and unmatched learners. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Cohort composition matters

Cohort composition matters is mainly about keeping the group definition stable. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is comparing the same students before and after rather than this year’s class with last year’s. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is using different cohorts as though they form a simple before-after design. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the population is explicit. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Teacher prompting can change with enthusiasm

Teacher prompting can change with enthusiasm is mainly about recognising implementation energy as part of the intervention context. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is the teacher unconsciously giving more cues during the new method’s first week. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is assuming only the named strategy changed. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that observation records meaningful shifts in support. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Novelty can temporarily raise engagement

Novelty can temporarily raise engagement is mainly about separating initial interest from durable learning. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is students working harder because the new routine is unfamiliar. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is treating one energetic week as permanent instructional superiority. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that follow-up evidence is collected after novelty fades. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Expectancy can influence interpretation

Expectancy can influence interpretation is mainly about precommitting to evidence to reduce confirmation bias. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is deciding in advance what would count as improvement and what would count as no change. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is looking hardest for evidence that justifies the teacher’s investment. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the inquiry includes disconfirming evidence. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Implementation fidelity must be checked

Implementation fidelity must be checked is mainly about knowing whether the intervention happened as designed. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is recording how often retrieval prompts were actually used. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is analysing a strategy that was only implemented sporadically. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that dose and adherence are documented. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Fidelity should not become rigidity

Fidelity should not become rigidity is mainly about allowing justified adaptation while recording it. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is shortening prompts after observing excessive lesson-time cost. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is quietly changing the intervention halfway through and analysing it as one fixed method. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that adaptations are dated and explained. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Process measures can explain outcome patterns

Process measures can explain outcome patterns is mainly about collecting evidence from the mechanism itself. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is tracking how many students retrieve without prompts before looking at test scores. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is using only the final mark. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that process evidence helps test why change occurred. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Error patterns can be more diagnostic than totals

Error patterns can be more diagnostic than totals is mainly about seeing whether the intended weakness changed. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is fewer method-selection errors after contrastive teaching. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is reporting only a one-point average gain. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the error type aligns with the intervention mechanism. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Student explanations can reveal changed reasoning

Student explanations can reveal changed reasoning is mainly about using qualitative evidence carefully. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is students more accurately explaining why two problems require different methods. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is treating interview comments as proof of population-level impact. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that qualitative evidence supports mechanism interpretation without replacing outcome data. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

A delayed check tests durability

A delayed check tests durability is mainly about looking beyond immediate performance. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is retesting the target capability two weeks later. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is measuring only on the final day of the intervention. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that learning survives beyond short-term performance. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

A changed task tests transfer

A changed task tests transfer is mainly about seeing whether the improvement travels. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is using new numbers, context and representation with the same underlying principle. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is repeating the practised items. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that transfer evidence is stronger than rehearsal. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

A control or comparison group can help but is not magic

A control or comparison group can help but is not magic is mainly about using comparison where practical while respecting selection differences. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is comparing two similar classes while acknowledging non-random assignment. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is calling any second class a perfect control. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that baseline differences and context remain visible. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Within-class comparison can be useful

Within-class comparison can be useful is mainly about using repeated measures when separate groups are unavailable. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is tracking the same skill across several baseline and follow-up points. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is treating one pretest and one post-test as a complete causal experiment. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that time series increases interpretive strength. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Single-case logic can guide local decisions

Single-case logic can guide local decisions is mainly about using repeated observations for one learner or class without generalising broadly. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is showing a stable baseline, intervention change and follow-up pattern. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is publishing one classroom pattern as a universal effect. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the claim stays local. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Reversal designs have limits in education

Reversal designs have limits in education is mainly about recognising when removing a helpful intervention is impractical or unethical. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is not withdrawing a safety or accessibility support merely to strengthen causal inference. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is treating experimental purity as more important than learners. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that design strength is balanced with educational responsibility. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Statistical significance is not the same as educational significance

Statistical significance is not the same as educational significance is mainly about asking whether the change matters in practice. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is a small reliable gain that does not alter independent performance. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is celebrating a p-value without considering magnitude or consequence. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the effect is interpreted in educational terms. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Effect size is not enough without construct clarity

Effect size is not enough without construct clarity is mainly about keeping the meaning of the outcome visible. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is a large gain on narrow rehearsed items versus a modest gain on fresh transfer items. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is preferring the larger number automatically. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the outcome’s educational meaning guides judgment. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Subgroup patterns should be treated cautiously

Subgroup patterns should be treated cautiously is mainly about checking whether the intervention helps different learners differently. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is novices benefiting while advanced learners show little change. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is declaring a subgroup effect from tiny numbers. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that subgroup claims require enough evidence. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Negative results can improve teaching

Negative results can improve teaching is mainly about allowing no-change findings to guide revision. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is discovering that the new prompt did not improve independent transfer. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is changing the method instead of hiding the result. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that inquiry values information over vindication. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Mixed results can identify boundary conditions

Mixed results can identify boundary conditions is mainly about learning where an intervention works and fails. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is immediate recall improving while transfer remains unchanged. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is calling the whole intervention either success or failure. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that different outcomes are interpreted separately. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Burden belongs in the evaluation

Burden belongs in the evaluation is mainly about considering teacher time and learner effort alongside gains. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is a routine producing small gains but consuming fifteen minutes every lesson. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is ignoring opportunity cost. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that benefit is weighed against displaced learning. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Access and inclusion belong in the evaluation

Access and inclusion belong in the evaluation is mainly about checking whether the change creates barriers. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is a fast oral routine disadvantaging students who need more processing time. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is averaging results without examining participation. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that implementation quality includes legitimate access. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

External research expands alternative explanations

External research expands alternative explanations is mainly about using prior studies to inform rather than dictate local interpretation. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is reading evidence on retrieval, cognitive load or formative assessment before revising the inquiry. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is treating one classroom as isolated from the research base. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that local evidence and external knowledge constrain each other. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

A causal claim should have a confidence level

A causal claim should have a confidence level is mainly about using language proportional to design strength. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is saying the pattern is consistent with the intervention contributing to improvement. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is writing ‘proved’ after a simple before-after comparison. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that wording reflects the evidence. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Illustrative data must be labelled as illustrative

Illustrative data must be labelled as illustrative is mainly about never inventing student results for a public case study. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is using a fictional 12/20 to 16/20 example clearly marked hypothetical. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is writing hypothetical numbers as though collected from a real class. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that readers can distinguish teaching illustration from empirical result. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Stop, adapt or continue should be pre-planned

Stop, adapt or continue should be pre-planned is mainly about turning inquiry into a decision process. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is stopping if burden is high and transfer remains flat, adapting if process improves but independence does not, continuing if durable transfer strengthens. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is continuing because effort has already been invested. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that decision rules are stated before sunk cost dominates. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

A later return check protects against temporary effects

A later return check protects against temporary effects is mainly about revisiting the question after ordinary conditions resume. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is checking the skill one month later after the intervention is no longer novel. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is closing the inquiry immediately after the best result. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the improvement remains visible over time. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Replication within the same classroom strengthens confidence

Replication within the same classroom strengthens confidence is mainly about trying the mechanism on another topic or cycle. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is using the same feedback routine in a second writing unit. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is treating one successful unit as permanent proof. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the pattern recurs under a new content context. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

Replication across classrooms changes the claim

Replication across classrooms changes the claim is mainly about requiring stronger coordination before generalising beyond one teacher. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is several teachers using a shared protocol with comparable evidence. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is calling one teacher’s result school-wide evidence. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the scope of the claim expands only with the evidence. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

The final purpose is a better next teaching decision

The final purpose is a better next teaching decision is mainly about keeping inquiry tied to action rather than publication-style certainty. Teacher inquiry is strongest when the mechanism is written down before the result arrives. What exactly changed in the lesson? What should improve if that change matters? Which other conditions could produce the same observed pattern? These questions make later interpretation less vulnerable to enthusiasm and hindsight.

A practical example is retaining the routine, adapting it or returning to the prior method based on evidence. The example matters because classrooms are open systems. Students continue learning, teachers continue adapting, assessments vary and school life introduces events that cannot be held constant. Local inquiry therefore aims for disciplined explanation, not impossible experimental purity.

The common failure is collecting data with no decision in mind. This usually creates a story with one cause because one cause is easier to communicate. But the educational value of inquiry comes from preserving enough alternatives to decide what to test next. A teacher can act under uncertainty without pretending the uncertainty disappeared.

The acceptance test is that the inquiry ends with a justified next move and a plan to recheck. Evidence should connect the intervention to the mechanism and the outcome: comparable task, documented support conditions, process signal, fresh or delayed performance and at least one serious alternative explanation. The stronger that chain becomes, the stronger the causal language may become.

A fully illustrative classroom example

Hypothetical only: a teacher notices that students choose the wrong method on mixed Mathematics problems. For three lessons the teacher introduces contrastive pairs and asks students to explain which cue selects each method. Before the change, an illustrative class average on a fresh mixed set is 12/20. One week later, an illustrative average on another mixed set is 16/20.

The first conclusion is only that performance improved on the later set. Before attributing the gain to contrastive teaching, compare item demand, content balance and support; check whether the students had already seen similar problems; inspect whether the teacher gave more method cues; look at method-selection errors specifically; and use a delayed fresh set. If selection errors fall while arithmetic errors stay similar, support remains constant and the improvement survives new items after delay, the mechanism story becomes more credible.

Even then, the responsible wording may be: The pattern is consistent with the contrastive routine contributing to improved method selection in this class. That statement is useful. It is also more defensible than saying the routine proved effective for all students or all Mathematics topics.

A teacher-inquiry causal checklist

  • Intervention: What exactly changed?
  • Mechanism: Why should that change affect this capability?
  • Prediction: What specific pattern should appear if the mechanism is right?
  • Baseline: Do I have more than one observation where possible?
  • Task: Are pre/post tasks comparable in construct and demand?
  • Support: Did prompts, notes or teacher help change?
  • Exposure: Did students practise the same items or format?
  • Co-changes: What else changed during the period?
  • Process evidence: Did the hypothesised behaviour itself change?
  • Delay: Does improvement survive time?
  • Transfer: Does it survive a fresh or changed task?
  • Burden: What time or opportunity cost did the intervention create?
  • Access: Did all intended learners have a fair opportunity to participate?
  • Alternative explanation: What else could plausibly explain the result?
  • Decision: Continue, adapt, stop or investigate further?
  • Return: When will I check the pattern again?

Frequently asked questions

Can a teacher ever prove an intervention worked?

Strong causal evidence is possible with suitable designs, but ordinary classroom inquiry often supports more bounded conclusions. The goal is a defensible local decision, not always proof at research-trial strength.

Is a pretest and post-test enough?

It is useful but weak for causal inference because task difficulty, practice effects, regression to the mean and other changes can explain part of the gain. Additional baseline, fresh tasks and process evidence help.

What is regression to the mean in classroom inquiry?

When an intervention begins after an unusually low score, later scores may rise partly because extreme measurements tend to be followed by less extreme ones. Multiple baselines reduce this risk.

Do I need a control class?

Not always. A comparison class can help, but non-random groups have their own differences. Repeated within-class evidence, changed tasks and mechanism measures can still improve local inference.

What is implementation fidelity?

It is evidence about whether the intervention was actually delivered as intended. If the routine occurred only twice instead of ten times, the interpretation of a weak result changes.

Should teachers avoid adapting an intervention?

No. Adaptation is often necessary. Record what changed and when so the inquiry does not pretend one stable intervention was delivered throughout.

What if immediate scores improve but delayed transfer does not?

Then the intervention may have improved short-term performance or familiarity without producing the durable, flexible learning you wanted. Adapt the design rather than averaging the two outcomes.

Can student comments count as evidence?

Yes, as qualitative evidence about experience, reasoning or implementation. They should not be treated as a substitute for performance evidence when the claim concerns capability.

How should I report hypothetical examples?

Label them explicitly as illustrative or fictional. Never present invented numbers as student results.

What is a good final sentence for a small teacher inquiry?

Use proportionate language: what changed, under which conditions, what alternative explanations remain, and what the next check will be. Avoid universal claims from one classroom cycle.

The quiet conclusion: inquiry is the discipline of not falling in love with the first explanation

Teaching improves through action. Teachers have to try things before perfect evidence exists. Inquiry does not ask them to stop acting until every variable is controlled. It asks for something more practical: make the change visible, state what should happen, compare fairly, look for the mechanism, keep alternatives alive and return later to see whether the result survives.

A teaching change earns confidence gradually. First the classroom feels different. Then process evidence changes. Then comparable performance improves. Then the improvement survives delay, reduced support and a fresh task. Each step narrows the alternative explanations. That is enough to make teacher inquiry powerful without pretending an ordinary classroom is a laboratory.