Working Data skill
Measure event outcomes
Evaluate what an event changed using reconciled participation, experience, behavior, and follow-up evidence.
Procedure
- Restate intended outcomes, target groups, baseline, time horizon, and decisions.
- Define participation, completion, reach, access, experience, learning, connection, action, and longer-term indicators.
- Specify numerator, denominator, source, collection timing, owner, coverage, and privacy limits.
- Reconcile registration, attendance, session, stream, survey, support, finance, sponsor, and follow-up data.
- Keep no-shows, partial participation, blocked tracking, missing responses, duplicates, and unknowns visible.
- Segment only where groups are large enough and the distinction informs action.
- Compare evidence with baseline, target, qualitative accounts, cost, and plausible alternative explanations.
- Publish bounded findings, limitations, decisions, actions, and later review.
Guardrails
- Do not claim causation from attendance, impressions, satisfaction, or self-selected testimonials.
- Never identify participants through small-group reporting or joined datasets.
- Do not change definitions after seeing results without preserving both versions.
--- name: measure-event-outcomes category: data description: Evaluate what an event changed using reconciled participation, experience, behavior, and follow-up evidence. Use when organizers need more than attendance and satisfaction counts. --- # measure-event-outcomes ## Procedure 1. Restate intended outcomes, target groups, baseline, time horizon, and decisions. 2. Define participation, completion, reach, access, experience, learning, connection, action, and longer-term indicators. 3. Specify numerator, denominator, source, collection timing, owner, coverage, and privacy limits. 4. Reconcile registration, attendance, session, stream, survey, support, finance, sponsor, and follow-up data. 5. Keep no-shows, partial participation, blocked tracking, missing responses, duplicates, and unknowns visible. 6. Segment only where groups are large enough and the distinction informs action. 7. Compare evidence with baseline, target, qualitative accounts, cost, and plausible alternative explanations. 8. Publish bounded findings, limitations, decisions, actions, and later review. ## Guardrails - Do not claim causation from attendance, impressions, satisfaction, or self-selected testimonials. - Never identify participants through small-group reporting or joined datasets. - Do not change definitions after seeing results without preserving both versions. ## Done - An outcome report defines measures, populations, sources, coverage, and uncertainty - Reconciliation checks explain missing, duplicate, partial, and conflicting evidence - Findings, limitations, actions, owners, and follow-up dates are recorded