AC9.3 · learner packet

Cooperation
and conflict.

Harbour Commons protocol and five records for testing claims without turning learners into the experiment.

No real stakes. No labels.

Tokens have no connection to grades, money, food, privileges, or access. Only anonymous group totals are recorded. Participant, observer, analyst, recorder, and independent-study routes are equivalent.

Learner
Group code or independent route
Start date

Final portfolio

Preregistered question and method Anonymous dataset and graph Lab-style report or equivalent Reproducibility and revision record

Fictional simulation · Weeks 3–5

Harbour Commons

A coastal growing cooperative draws water from a shared fictional reservoir. The exercise observes aggregate patterns under different rules; it does not identify winners, losers, or personality types.

Starting stock

fictional water tokens at the start of every condition

Private action

0 · 2 · 4 · 6

tokens requested per role card and round

Renewal

½ remains

Add half the remaining stock after withdrawals, never exceeding 60

Round rule

Ending stock = starting stock − aggregate requests. If requests exceed stock, record unmet demand and set stock to zero. Renewal = half the ending stock, rounded down. Next stock = ending stock + renewal.

Condition A · rounds 1–3

Baseline

No discussion. Private action slips. Publish group requests, unmet demand, and stock only.

Condition B · rounds 4–6

Communication

Two minutes of discussion before each round. Agreements are not enforced.

Condition C · rounds 7–9

Self-made rule

The group writes one allocation or replenishment rule plus a review process.

Condition D · rounds 10–12

Unequal capacity

Fictional role cards change stated needs. Cards are random and never tied to learner identity.

Safety and validity rules

Reset stock to 60 before every new condition.
Record anonymous group totals only; destroy private action slips after totals are checked.
No real reward, penalty, deception, forced disclosure, or public individual comparison.
Stop a round if privacy fails, instructions change mid-round, or a learner withdraws consent.
Record deviations and missing data instead of quietly repairing them.

Tool 01 · Week 3

Question and preregistration

Commit to a narrow comparison before seeing the results. A changed prediction remains visible in the revision record.

Research question: how does one specified condition relate to one measured outcome? Hypothesis and predicted direction Reasoning or source supporting the prediction

Changed variableWhat differs between conditions?Measured outcomeWhat aggregate result will be recorded?Held constantWhat must remain the same?Possible confoundWhat else might explain a difference?

Missing-data ruleStop condition

What result would challenge or weaken the hypothesis?

Protocol versionDate fixedIndependent reproduction checker

Tool 02 · Weeks 1–5

Source evidence audit

Use one copy per source. Do not treat a historical case, mathematical model, laboratory experiment, and field study as the same kind of evidence.

Author and sourceEvidence type

Claim in your own words What was observed, modelled, measured, or reported?

Key assumptionsSample, case, or setting

Alternative explanation or conflicting evidence What can this evidence travel to—and where must the conclusion stop?

Confidence in the narrow claim Low Medium HighWhy?

Tool 03 · Week 4

Anonymous round log

Record group totals and protocol quality. Keep private actions out of the dataset and portfolio.

RoundConditionStartGroup requestUnmetEnd + renewalWork / deviation

Self-made rule used in Condition CFictional capacity change in Condition D

Data cleaning notes: missing values, corrected transcription, exclusions, and reasons

Graph space

Label condition, outcome, scale, units, and missing data. A line between rounds does not itself prove causation.

Tool 04 · Weeks 1 and 6

Claim ladder

Each rung needs more support than the one below it. Do not climb by changing vocabulary alone.

01 · Recorded observation

What appears directly in a source or anonymous dataset?

Confidence

02 · Narrow inference

What relationship might explain the observation here?

Alternative

03 · Mechanism explanation

What process is proposed, and what evidence tests it?

Assumption

04 · Transfer claim

Where else might the finding apply, and what differs there?

Limit

05 · Political conclusion

What should people do—and what additional values or evidence does that require?

Additional support needed

A simulation result cannot prove:

that people are naturally good or selfish; that one institution is universally legitimate; that punishment is just; or that a political identity must be adopted.

Tool 05 · Week 6

Report and revision planner

Make the method reproducible, the evidence inspectable, and the limits impossible to miss.

Question and hypothesisWhat was predicted, and why?MethodProtocol, variables, privacy, and deviationsResultsPattern, summary, graph, and uncertaintyDiscussionInference, sources, and alternative explanationsLimitsSample, model, measurement, and transferPolitical reflectionConclusion plus values or evidence beyond the simulation

Reproducibility check

What another person reproducedAmbiguity or mismatch found

Substantive revision after challengeWhy the revision improves the work

Final review

Raw data and corrections remain visible. Observation and inference are separate. A serious alternative explanation is addressed. No individual action or identity is disclosed. The political conclusion names its additional support needs.

Cooperationand conflict.