Use this working method to prepare a specific, testable project. It is an editorial planning guide, not an eligibility decision or a promise of funding.
Participation → Response → Change → Review
Separate activity from outcomes
Visits, downloads and submitted reports show activity. They do not show that a public service improved. Describe the intended change: an acknowledged request, a documented response, an accessible decision or a resolved issue.
Write an evidence chain
Link an activity to an expected response and then to a change that matters to participants. For each link identify an evidence source and an assumption. An assumption that an institution will respond needs testing rather than a higher target.
Set a baseline and a denominator
Record the starting situation and who is eligible to participate. Report totals with a period and a denominator. For example, ten acknowledged requests out of twenty submitted in a defined test is clearer than a claim of high engagement.
Look for exclusion
Check which communities, languages, ages and connectivity conditions are missing from the results. Collect only the information needed for this analysis. Avoid publishing small groups in a way that makes individuals identifiable.
Review mixed evidence
Combine a few transparent counts with participant feedback, staff observations and a record of institutional responses. Note failures and alternative explanations. Improvement can reflect many influences; avoid claiming that your tool alone caused a broader policy change.
Keep a decision log
At each review record what you observed, what changed in the design and what the next test will answer. Share a readable summary with participants. A dashboard is useful when it supports a decision; it is not an outcome by itself.
Keep the same cohort when reporting a response rate
An outcome record needs a starting point, an eligible group and a defined observation window. This invented civic-service example separates requests submitted from responses documented. It is not a grant rule, a funded project or an evaluation of an actual institution.
Define the unit before collecting counts: one eligible request with a stable non-identifying reference. State what counts as a response and how long the observation window lasts. A click or automatic receipt is not a substantive response.
Take a baseline using the same definition. Record exclusions and unfinished cases, not just successes. Keep personal details outside a public worksheet and avoid publishing small groups that identify individuals.
At review, match the same cohort. Do not divide responses to older cases by new requests from the current week. Report the numerator, denominator and unresolved cases together.
Use the result to choose a next action, while retaining alternative explanations. A higher documented response rate does not by itself show that a civic tool caused a policy change or improved every participant’s experience.
20 eligible requests, 8 substantive responses within the defined window: 8/20 = 40%.
Later cohort — fictional
20 different eligible requests, 12 responses under the same rule: 12/20 = 60%. The remaining 8 are unresolved at review.
Comparison
Difference: 20 percentage points, not a 20% relative increase. Relative change is (60−40)/40 = 50%, if that comparison is justified.
Limits
Different case mix, staff availability or reporting practice could explain the difference. No causal result has been established.
Working record
Unit __; eligibility __; cohort/window __; baseline __; response rule __; numerator/denominator __; unresolved __; alternative explanation __; next decision __.
Agenda 2063 supplies public-governance context, not a rule for this exercise. Check the exact conditions of any real funding call separately; no open application window is asserted here.
