Direct evidence
Supported by data, documents, testimony, measurement or another identifiable source.
The Civic Method
A rigorous structure for learning through policy.
The Civic Method connects public purpose, evidence, choices, implementation and evaluation within one continuous decision process.
Six stages
Questions before instruments.
Each stage is defined by the decision question it must answer and the evidence artifact it leaves behind.
- 01
Define
What public value are we trying to create, and for whom?
Typical outputDecision mandate
Clarify authority, public purpose, affected groups, constraints and the decision that must be made.
- 02
Understand
What system, institutions, incentives and lived realities shape the problem?
Typical outputEvidence and system map
Combine evidence forms, examine the problem boundary and identify where knowledge is strong, contested or absent.
- 03
Formulate
Which mechanisms could produce the intended change?
Typical outputTheory of change
Connect interventions to outcomes through explicit, contestable causal assumptions.
- 04
Compare
What are the consequences, trade-offs and implementation requirements of each option?
Typical outputPolicy options appraisal
Compare credible alternatives, including distributional effects, administrative feasibility and the option not to intervene.
- 05
Test
Which assumptions can be piloted, simulated or stress-tested before scale?
Typical outputTesting protocol
Select proportionate methods with ethical safeguards, decision thresholds and clear limits on inference.
- 06
Learn
What evidence will determine whether the policy should continue, adapt or stop?
Typical outputEvaluation and learning framework
Define indicators, feedback paths, review moments and responsibility for acting on what is learned.
Not a rigid sequenceThe method’s depth and form are proportionate to the decision, the evidence available and the consequences of being wrong. New evidence may return a team to an earlier stage.
Evidence standard
Claims should reveal what supports them.
Reasoned interpretation
The strongest explanation available, with the reasoning and credible alternatives made visible.
Necessary premise
A premise required to proceed but not yet verified, paired with a way to test or revisit it.
Material uncertainty
A gap that may change the decision and therefore needs a response, boundary or deliberate acceptance.
Institutional learning
Preserve the reasoning, not only the decision.
- Objective
- Evidence
- Assumptions
- Options
- Decision
- Observed outcomes
- Adaptation
Limits and ethics
Method does not replace legitimacy.
Public experimentation is constrained by law, democratic authority, proportionality, consent, equity and the possibility of harm.
CMI will not present technical analysis as a substitute for political or democratic judgment.
Testing should stop where potential harm cannot be acceptably bounded, observed or repaired.
Participation must not become symbolic extraction from affected communities.
Evaluation designs must disclose their limits and avoid causal claims that the method cannot support.
Working with CMI
Begin with the policy question.
CMI welcomes conversations with public institutions, research organizations and public-interest foundations confronting consequential policy questions.