What to Measure for Each MAR Hypothesis

Preventive Diplomacy

How long does it take someone with authority to acknowledge a verified concern?

Suggested outcome to record: Minutes from verification to acknowledgment, plus cases with no acknowledgment by the exercise deadline.
How to interpret it: Use the same cases and available information in each arrangement. Record premature escalations and missed concerns separately. An acknowledgment is evidence of contact. Log whether the participants actually clarified the issue.

Read the original hypothesis Review this hypothesis in the Lab

Peace Agreements

Can reviewers verify that a promised civilian service has resumed?

Suggested outcome to record: Number of named service milestones with dated, independently corroborated evidence out of all milestones due.
How to interpret it: Keep overdue, disputed, and unverified milestones separate. A service can be working while the parties disagree about who restored it or whether a commitment was fulfilled.

Read the original hypothesis Review this hypothesis in the Lab

Positive Peace

Can staff follow the agreed procedure after the people in charge change?

Suggested outcome to record: Required steps completed with documented authority, out of all steps due in the handover exercise.
How to interpret it: Use the same cases and deadlines before and after a staff handover. Look for tasks that only one person can complete. A smoother handover establishes a narrow process result; broader institutional trust needs separate evidence.

Read the original hypothesis Review this hypothesis in the Lab

Economic Interdependence

Can partners keep essential shipments moving through a disruption?

Suggested outcome to record: Confirmed civilian units received by a stated deadline, unmet demand, and cost per participant in the same period.
How to interpret it: Hold the volume requested, shipment dates and delivery resources steady when testing a coordination change. Report the delivery ceiling, goods actually accepted, and any participant newly dependent on a single supplier as distinct measures.

Read the original hypothesis Review this hypothesis in the Lab

Arms Control

Can participants verify a permitted civilian notification without spreading a disputed report?

Suggested outcome to record: Verified civilian notices acknowledged by the deadline, plus duplicates, corrections and disputed reports.
How to interpret it: Use fictional notices and approved public information in an exercise. Preserve disagreements in the record. This measures the handling of civilian notices; established arms-control processes have their own authorities and evidence.

Read the original hypothesis Review this hypothesis in the Lab

Humanitarian Cooperation

Does a verified message help authorized assistance reach the people it is meant for?

Suggested outcome to record: Time from verification to authorized dispatch, plus usable supplies actually received by the deadline.
How to interpret it: Record access, staff and supplies available in each case. Track arrival separately from dispatch and avoid counting the same delivery twice. Record privacy incidents and people or areas the exercise could not cover.

Read the original hypothesis Review this hypothesis in the Lab

Sovereign Autonomy

Which essential services can each participant keep running if it leaves the shared arrangement?

Suggested outcome to record: Essential functions with a tested local fallback at a stated duration and service level.
How to interpret it: Test each participant separately. Count a function as unknown if the local replacement was never tested. Record the staffing, permissions and effort required to use the fallback.

Read the original hypothesis Review this hypothesis in the Lab

Shared Cyber Defense

Can a civilian operator independently confirm a defensive report before acting on it?

Suggested outcome to record: Minutes to independent verification, known false alerts, known missed cases, and unresolved reports.
How to interpret it: Use synthetic or authorized aggregate records. Compare equal reviewer time and the same verified cases. The exercise tests review quality and operational continuity; it does not attribute an intrusion.

Read the original hypothesis Review this hypothesis in the Lab

Energy Continuity

Do priority civilian services recover within their target time?

Suggested outcome to record: Minutes until the required service is verified as working, plus missed service hours during the interruption.
How to interpret it: Start both clocks at the same event and use the same definition of restored service. Record local fallback and any added staff or equipment. A recovered service may still have lost records or failed to cover all users.

Read the original hypothesis Review this hypothesis in the Lab

Food Security

Does earlier warning give a food distributor enough time to protect supply?

Suggested outcome to record: Hours from independently verified warning to the decision deadline, and usable food delivered within the window.
How to interpret it: Use confirmed usable stock, delivery constraints and the same demand in every case. Compare lead time and actual supply separately. An early warning can still leave a shortage or prompt an unnecessary response.

Read the original hypothesis Review this hypothesis in the Lab

Water Cooperation

Do the participants have timely, comparable observations to work from?

Suggested outcome to record: Required periods and places with a current observation taken under the agreed definition.
How to interpret it: State the reporting period, measurement definition and age limit. Show missing locations and conflicting records. Better coverage can support a shared discussion; it does not decide water rights or future water availability.

Read the original hypothesis Review this hypothesis in the Lab

Public Health

Can authorized teams act on a verified civilian notice while protecting sensitive information?

Suggested outcome to record: Time to acknowledgment and confirmed receipt of requested supplies, plus any prohibited disclosure.
How to interpret it: Work with aggregate or fictional data and approved recipients. A fast alert alone leaves the supply result unknown. Clinical or population health outcomes require separate evidence and study design.

Read the original hypothesis Review this hypothesis in the Lab

Climate Adaptation

Did a forecast help more than a simpler benchmark for the same place and date?

Suggested outcome to record: Typical absolute error in the original unit, plus the number of matched observations and their forecast lead times.
How to interpret it: Compare forecasts issued before the event and evaluate both on the same observations. Show missing cases and the definition of the benchmark. A lower average error may hide poor performance during rare events.

Read the original hypothesis Review this hypothesis in the Lab

Disaster Preparedness

Did the intended people receive and acknowledge a useful warning in time?

Suggested outcome to record: Share of intended recipients with confirmed acknowledgment by the deadline, plus missed or false alerts.
How to interpret it: Specify the audience and accessible channels before testing. Keep verified receipt, acknowledgment and authorized response separate. Assess people reached through local channels alongside the shared process.

Read the original hypothesis Review this hypothesis in the Lab

Information Integrity

Did reviewers trace a disputed claim and correct the record when evidence changed?

Suggested outcome to record: Traceable independent sources, time to a documented correction, and unresolved disputes.
How to interpret it: Trace duplicated reports to the originating observation. Record when a correction was issued and acknowledged. A correction improves the documented record without settling every disagreement.

Read the original hypothesis Review this hypothesis in the Lab

Digital Sovereignty

Can local staff continue essential work without the shared software or outside provider?

Suggested outcome to record: Essential functions retained at the required service level during a stated period of local operation.
How to interpret it: Use a tested local copy, named responsibilities and actual staff coverage. Include time to switch, missing features, and services that can no longer be operated. Keep results for different participants visible.

Read the original hypothesis Review this hypothesis in the Lab

Scientific Diplomacy

Can each participant independently reproduce the same technical result?

Suggested outcome to record: Independent reproduction attempts meeting an agreed tolerance, out of all attempts due by the deadline.
How to interpret it: Set the tolerance and protocol before seeing results. Track who had access to the inputs and whether they shared a tool or original source. Agreement on one test result says little about the broader relationship.

Read the original hypothesis Review this hypothesis in the Lab

Supply Chain Resilience

How much stock is needed before confirmed deliveries can meet demand?

Suggested outcome to record: Minimum usable opening stock under a dated supply schedule, and the first period with unmet demand.
How to interpret it: Count only receipts that can be used in the relevant period. Stress the arrival date, daily demand and local delivery limit separately. A stock estimate changes when those assumptions change.

Read the original hypothesis Review this hypothesis in the Lab

Anti-Corruption & Rule of Law

Can a reviewer tell who authorized a consequential action and why?

Suggested outcome to record: Material actions with a traceable approval and complete decision record, out of all material actions reviewed.
How to interpret it: Check who was responsible, what was approved, what changed, and whether someone could challenge the record. A traceable decision still requires human oversight of the facts and authority.

Read the original hypothesis Review this hypothesis in the Lab

Strategic Deterrence

Can civilian services continue when cooperation is withheld or interrupted?

Suggested outcome to record: Civilian functions with a verified local fallback, measured at a stated service level and time.
How to interpret it: Test civilian continuity using synthetic or authorized aggregate records. The results can show whether a particular dependency matters under those assumptions. They provide no estimate of coercion, escalation, or conflict outcomes.

Read the original hypothesis Review this hypothesis in the Lab

What to measure for each MAR hypothesis

Preventive Diplomacy

Peace Agreements

Positive Peace

Economic Interdependence

Arms Control

Humanitarian Cooperation

Sovereign Autonomy

Shared Cyber Defense

Energy Continuity

Food Security

Water Cooperation

Public Health

Climate Adaptation

Disaster Preparedness

Information Integrity

Digital Sovereignty

Scientific Diplomacy

Supply Chain Resilience

Anti-Corruption & Rule of Law

Strategic Deterrence