A student can identify a weak argument yet still struggle to decide what to do next. Knowing that a school rule has costs, a public claim is incomplete or a technology can create dependency is only the beginning. Responsible judgement also requires the ability to compare real alternatives, explain which consequence follows from which assumption, act within appropriate authority and update a decision when new evidence arrives. Those skills can be practised in a structured decision laboratory.
A Ministry of Education V3.0 decision laboratory is a classroom or institutional learning exercise that lets participants test a realistic question without exposing real people to unnecessary risk. It combines evidence checking, causal maps, alternative routes, opportunity costs, uncertainty, counterfactual reasoning and an after-action review. The value is not theatrical simulation. It is disciplined practice that produces a decision record another person can inspect, question and improve.
This independent eduKateSG education-design reference is not an official Singapore Ministry of Education programme, policy simulation system or approval for student experiments on real communities. The examples below use fictional or safely authorised information. Children must not conduct harmful interventions, take real financial risks, impersonate public officers or negotiate away their own or others’ rights.
1. What Makes a Decision Laboratory Different From a Debate?
A debate can develop argument, listening and response. A decision laboratory adds practical accountability. The participants must specify what is being decided, who may decide, which conditions must be respected, what evidence is available, which option they choose and what result would make them revise the choice. A persuasive speaker does not win merely by sounding confident.
Consider a school considering longer library opening hours. One group says longer hours support students; another says staffing costs are too high. Both may be reasonable perspectives. The laboratory asks what problem actually needs solving, which learners would benefit, what alternative resources exist, whether a pilot is feasible, what the added workload is, who can authorise the change and how to check whether more students gain usable study access.
The OECD’s Anticipation–Action–Reflection framework frames learning as a cycle in which people consider future consequences, act responsibly and reflect on results. A V3.0 decision lab is one practical educational interpretation of that learning cycle. It does not imply that the OECD has endorsed this specific classroom procedure.
2. Six Required Parts of a Sound Decision Lab
| Part | Guiding question | What the learner produces |
|---|---|---|
| Problem | What specific condition needs improvement? | Bounded decision statement |
| World and rules | Who is affected, who can act and what constraints apply? | Roles, facts and authority map |
| Evidence | What supports each important claim? | Source-quality and uncertainty record |
| Options | What feasible alternatives exist, including a status-quo option? | Comparable route table |
| Action | Which route is proportionate under the stated conditions? | Decision, owner and safe test |
| Return | What result or new fact would change the next decision? | Review, repair and transfer note |
The laboratory should not reward novelty over accuracy. A well-supported decision to retain an existing arrangement may be better than an exciting reform that creates avoidable harm. Equally, students should be willing to recommend change when evidence shows that the current route leaves a serious problem unresolved.
3. The Twelve-Step Lab Design and Delivery Procedure
Step 1 — Select an educationally legitimate case
Choose a topic connected to existing subject outcomes, such as interpreting a resource budget in Mathematics, comparing information sources in English, evaluating a science experiment or examining a transport access scenario in Geography. Ensure the material is age-appropriate and does not require students to reveal private experiences.
Step 2 — Write the precise decision boundary
Specify what participants are allowed to recommend. For example: ‘Choose a one-term trial for distributing revision materials in a fictional school.’ Do not imply that students can decide actual teacher staffing, disciplinary matters or public policy simply because the exercise simulates a committee.
Step 3 — Define the baseline and intended improvement
Provide the existing arrangement and an observable problem. A library may close before some students finish their approved activities, or a platform may have uneven access. State which evidence the simulated institution already possesses and which information is not available. A decision without a baseline cannot be evaluated meaningfully.
Step 4 — Issue a fair evidence packet
Give all participants the same essential facts, relevant rules and plausible limitations. Sources should be identified as real, fictional or simplified. Do not present invented statistics as national facts. Where a genuinely contested question is involved, explain what is known and what remains a judgement of values or priorities.
Step 5 — Map stakeholders and legitimate authority
Ask who benefits, who bears the cost, who has necessary expertise and who can approve or stop the choice. Include the learner perspective without pretending that children hold adult legal or safeguarding duties. A stakeholder map should help participants reason rather than turn real people into targets of speculation.
Step 6 — Develop at least three feasible options
Include the current arrangement when it is a genuine choice, a modest or reversible intervention, and a credible alternative design. Each option must be capable of meeting the problem under plausible conditions. Avoid comparing the preferred option against deliberately absurd alternatives.
Step 7 — Separate causal claims from predictions
A route map shows how a change might produce a consequence. Students must mark which links are observed, supported by research, inferred or unknown. A plausible mechanism is useful but is not proof of an effect. The learner should be able to state the evidence that would test the most important missing link.
Step 8 — Ask the counterfactual question
What would likely have happened if the chosen action had not occurred, or if another option had been used? This does not mean inventing an alternative history as fact. It means building a disciplined comparison under stated assumptions. Where the scenario cannot establish the counterfactual, students must openly acknowledge the uncertainty.
Step 9 — Compare hidden costs and unfair distribution
Count teacher time, family burden, accessibility, learning continuity, privacy, resources and delayed consequences where relevant. A programme can improve average outcomes while making participation harder for a specific group. The lab should examine distribution rather than hide it behind a single convenient average.
Step 10 — Make a reasoned provisional decision
Students select an option and present an evidence-based explanation, including its limits. A good recommendation identifies what will happen first, who has authority, the condition under which the choice should stop and which groups need specific safeguards. Do not equate a confidently expressed conclusion with a well-reasoned one.
Step 11 — Introduce a new evidence or constraint card
After the initial decision, provide a previously unseen but plausible observation: the intervention costs more than expected, a subgroup lacks access, a core assumption proves false or an alternative unexpectedly performs better. Students must update the decision record and explain which argument changed. This is one of the most valuable parts of the laboratory.
Step 12 — Return to actual learning and transfer
Ask students to distinguish what the fictional exercise established from what a real institution would still need to investigate. Then present a second, different case without the original scaffolds. The objective is not memorising the first solution, but transferring a disciplined decision process to a new problem.
4. Counterfactuals Without Pretending to See Parallel Worlds
A counterfactual describes what might have happened under a different choice or condition. It is central to evaluating an intervention, yet rarely observable directly. We usually cannot see the same classroom both receive and not receive a programme at the exact same time under identical conditions. This is why comparison groups, suitable evaluation designs and cautious causal language matter.
| Statement | What it means | What is not established |
|---|---|---|
| Scores rose after a new method | A change occurred during the relevant period | That the method alone caused it |
| Similar groups showed different outcomes | The comparison may be informative | That every confounder was controlled |
| A route has a plausible mechanism | The proposed effect could happen under conditions | How large or likely the effect is |
| A simulation favoured one option | The model produced an outcome given assumptions | What will happen in all real schools |
| A later result contradicted the plan | An assumption or implementation may need review | That everyone who supported the first plan acted irrationally |
In a well-run lab, counterfactual reasoning disciplines claims rather than inflating them. The National Academies Framework for K–12 Science Education describes scientific practice as constructing and evaluating explanations, using models and revising conclusions through evidence. Students learn this intellectual habit when they recognise the limits of their own simulation.
5. Worked Lab: More Practice or Better Diagnosis?
Illustrative Mathematics scenario. A lower-secondary class shows persistent problems in algebra. Teachers are considering two changes: add the same worksheet volume for every learner, or provide diagnostic bridging lessons for learners whose fraction prerequisites remain weak. A third option is to keep the current approach temporarily while improving measurement.
Students receive fictional pre-intervention work samples, a fixed staff-time allowance and a statement that some learners struggle with symbolic representations while others are already ready for advanced questions. They compare how each option might use time, affect stronger learners and repair specific gaps.
One group’s recommendation might favour a diagnostic pilot with short independent follow-up tasks. Another may argue that the existing approach could be improved by adding targeted explanations without a separate programme. Both can receive high marks when they identify credible mechanisms, resources and uncertainty. The Secondary 1 Mathematics tutoring reference in Clementi provides a related diagnosis-first setting, while the lab remains a hypothetical policy exercise.
6. Worked Lab: A Library’s Opening Hours
A fictional school library opens until 3:30 PM. Some students say they need quiet space later, while the school has limited staffing. The lab’s stated goal is not to maximise hours at any cost: it is to increase usable, supervised study access while managing safety and resources.
The evidence packet includes fictional aggregated usage by hour, staffing constraints, accessible alternatives and a small student survey with clear sampling limits. Students compare a two-day extension, a supervised alternative study room and an improved booking system under current hours. A mid-exercise evidence card shows that some students cannot remain late because of transport.
The updated decision should consider whose access improves and whose does not. Students might preserve an afternoon option and add a daytime alternative, or recommend gathering better evidence before changing hours. They should not infer that one learner’s preference represents the whole school.
7. Worked Lab: A Public Notice and a Hidden Access Barrier
The lab gives students a fictional notice requiring families to register for an educational service online. The intended improvement is faster administration. A new information card reveals that some households cannot complete the digital identity step. The intervention must now be assessed not only for efficiency but for usable access.
Students can propose clearer instructions, an authorised alternative channel or staff support. Their decision receipt identifies the responsible institution and a later participation check. The exercise integrates language clarity, civic literacy, digital privacy and fairness without asking students to contact real public officers or submit actual personal data.
8. The Decision Lab Packet
| Packet component | Minimum contents | Reason |
|---|---|---|
| Case brief | Problem, learners, period and decision authority | Prevents endless, unbounded argument |
| Baseline | Current arrangement and observable gap | Anchors evaluation |
| Evidence cards | Cited facts or clearly marked fictional observations | Supports truthful analysis |
| Option cards | At least three credible routes | Prevents false binary choices |
| Cost cards | Time, access, money, privacy, workload and sustainability | Makes burdens visible |
| Constraint cards | Law, curriculum, safety and practical authority | Prevents unrealistic recommendations |
| New-evidence card | One material change after the first decision | Tests updating |
| Decision receipt | Choice, rationale, uncertainty, safeguards and review | Makes thinking inspectable |
9. A Forty-Five-Minute Lesson
| Minutes | Activity | What is collected |
|---|---|---|
| 0–6 | Read the problem and identify authority | Bounded decision statement |
| 6–13 | Annotate evidence and uncertainty | Evidence register |
| 13–21 | Map causes, alternatives and the counterfactual | Three-option route chart |
| 21–29 | Compare cost, fairness and risk | Trade-off table |
| 29–34 | Write an initial recommendation | Decision receipt |
| 34–40 | Reveal a new evidence card and revise | Change log |
| 40–45 | Identify the limits and transfer to a fresh micro-case | Independent reasoning note |
The sequence is adaptable. Younger students can use a simpler classroom-sharing problem, while upper-secondary and adult learners can compare richer information and quantitative assumptions. Do not make the simulated setting more complicated than the subject learning objective can support.
10. The Quality Rubric
| Criterion | Beginning | Proficient | Advanced |
|---|---|---|---|
| Problem framing | Repeats a broad slogan | Defines the actual decision and constraints | Recognises hidden assumptions about authority or need |
| Evidence | Uses impressions and selective examples | Identifies support, limitations and uncertainty | Tests alternative explanations and source scope |
| Counterfactual | Assumes the preferred path causes all gains | Compares against a credible alternative | Explains which causal links cannot yet be established |
| Ethics and distribution | Considers the primary user only | Includes other affected groups and safeguards | Revises options to reduce serious asymmetric burdens |
| Decision and return | States an opinion | Chooses a feasible action with an observable review | Updates decisively and proportionately after new evidence |
11. Avoid Five Failure Modes
- Simulation theatre: students perform roles but the decision has no measurable target or rules.
- Predetermined virtue: one politically or morally fashionable choice receives marks regardless of evidence.
- Fake causal certainty: invented numbers or model outputs are presented as results from real communities.
- Unfair complexity: reading load, disability-related barriers or inaccessible tools obscure the skill being assessed.
- No return: students make an initial recommendation but never confront a changed assumption or consequence.
The laboratory must also avoid exposing learners to personal financial obligations, social conflict, unsafe experiments or real disciplinary decisions. A controlled case is valuable precisely because it permits serious reasoning without putting people’s rights at risk.
12. The Institutional Implementation Loop
- Select a small set of cases aligned to existing subject learning outcomes.
- Use current references and label all constructed data explicitly.
- Train teachers to moderate evidence and debate fairly.
- Provide accessible alternatives to oral performance or complex digital interfaces.
- Observe how well students distinguish facts, assumptions and preferences.
- Use a new-evidence card to test genuine updating.
- Assess transfer with an unfamiliar case after an appropriate interval.
- Record weak parts of the case, repair them and maintain versions.
- Do not scale a teaching innovation solely because students enjoyed the simulation.
The V3.0 Pilots and Responsible Scaling reference describes how to test whether an educational intervention works across classrooms. The decision lab itself is a learning instrument, so it should be subject to the same scrutiny.
13. The Gold Standard
A student who can say ‘I would choose this route, given these assumptions; here is the strongest alternative; this group could be harmed; and this evidence would change my decision’ has learned more than the ability to win a debate. They have demonstrated a disciplined relationship between knowledge and action.
That is the practical promise of a decision laboratory: the learner rehearses intelligent responsibility in a safe setting, then becomes better equipped to recognise the boundaries and demands of real decisions.
Related Manuals and External Foundations
- MOE V3.0 — Complete Reference Library
- Reference 01 — How to Run a Route Audit
- Reference 02 — Hidden-Cost Ledger
- Reference 03 — Evidence and Claims Verification
- Reference 09 — Pilot Evaluation
- Reference 15 — Ethical Dialogue and Safeguards
- Reference 30 — Scientific Uncertainty and Forecast Literacy
- OECD — Anticipation–Action–Reflection Cycle
- National Academies — Scientific and Engineering Practices