After-Action Proportionality Review¶
Review process — instantiates Proportionality Calibration
Reviews whether the response produced intended safety, fairness, deterrence, repair, or learning without excessive collateral burden.
Every other mechanism in this family decides what to do; After-Action Proportionality Review is the only one that looks back and asks whether the response that was actually taken fit. It is a retrospective process: it reconstructs a closed case, compares the response that was executed against the harm it was meant to address and the burden it turned out to impose, and reports whether the match held. Its defining move is that it judges a decision already made — it produces no sanction, no band, no ladder of its own. What it produces is a verdict about fit (over-response, under-response, uneven application, or well-matched) and, crucially, the correction that verdict implies for the scales and thresholds the next case will run through. It is what turns proportionality from a one-time judgment into a feedback loop.
Example¶
A regional refinery's gas-detection system alarms on hydrogen sulfide near a processing unit. The shift lead, acting in seconds, orders a full site evacuation and shuts the unit down for eleven hours. Six weeks later the After-Action Proportionality Review convenes. It does not re-litigate the alarm; it asks whether the response fit what was knowable at 2 a.m. It reconstructs the decision-time picture: one sensor reading near threshold, no corroborating alarm, a crew of nineteen on site. Against that, it tallies the realized burden — the lost production, the restart hazards a cold shutdown introduced, the crew's growing habit of treating full evacuations as routine. Then it cross-checks: were the three comparable single-sensor alarms in the past year handled the same way? Two were; one was downgraded to a local sweep by a different lead. The review's output is not "the shutdown was wrong." It is "the response was defensible given the information, but our threshold for full shutdown versus local sweep is applied inconsistently across shifts, and the restart hazard is an uncounted burden — both go back to the escalation rubric." That correction, not a blame finding, is the point.
How it works¶
- Freeze the decision-time frame. Reconstruct what was actually known when the response was chosen, deliberately walling off everything learned afterward, so the review grades the choice, not the luck of the outcome.
- Tally the realized burden. Measure the cost the response actually imposed — on the target, on bystanders, on implementers, on trust — including burdens that never appeared in the original plan.
- Cross-check for even application. Line the case up against comparable past cases and look for unexplained divergence in how strongly similar situations were met.
- Route the correction. Convert each mismatch into a specific adjustment to the severity rubric, response scale, or thresholds the next case will use — the review's deliverable is the fix, not the finding.
Tuning parameters¶
- Review trigger — which cases get reviewed: all of them, a random sample, or only the high-burden and contested ones. Broader coverage catches quiet drift but costs reviewer time.
- Lookback window — how long after closure the review runs. Sooner keeps memory fresh; later lets downstream burden and repeat-offense data mature.
- Reviewer independence — whether the original decider sits on the panel. Independence reduces self-justification but loses context only the decider holds.
- Decision-time vs. outcome framing — how hard the process insists on judging by what was known then rather than how it turned out. The stricter the framing, the more it resists rewarding lucky over-reactions.
When it helps, and when it misleads¶
Its strength is that nothing else closes the loop: over-response, under-response, escalation that never stepped back down, and quiet inconsistency between deciders are all invisible until something looks back and names them. Because its deliverable is a routed correction, it is the mechanism that lets a response scale actually improve rather than ossify.
Its central failure mode is outcome bias — grading the decision by how the case ended rather than by whether the response fit what was known at the time.[n1] A shutdown that turned out unnecessary gets condemned; a lax response that happened to avoid disaster gets praised; both lessons are wrong. The classic misuse is running the review as a search for someone to blame, which teaches everyone downstream to over-respond defensively so no future review can fault them. The guarding discipline is to freeze the decision-time frame before any outcome is discussed, and to direct every finding at the rubric rather than the person.
How it implements the components¶
proportionality_review— this is the component: a structured after-the-fact check that the executed response stayed matched to severity, necessity, burden, and outcome.burden_assessment— it measures the burden the response actually imposed, including uncounted collateral costs, which is the input the fit verdict weighs against the harm addressed.equity_cross_check— comparing the case against similar past cases surfaces uneven application across deciders, times, or groups.
It defines no severity_assessment rubric, response_scale, or escalation_threshold of its own — those are set prospectively by mechanisms like the Response Matrix and the Moderation Strike System; this review consumes their outputs and feeds corrections back to them.
Related¶
- Instantiates: Proportionality Calibration — it is the review invariant made operational, closing the calibrate-act-review loop.
- Consumes: Response Matrix and the other selection mechanisms supply the executed decisions whose fit this review audits.
- Sibling mechanisms: Disciplinary Guideline · Moderation Strike System · Proportional Remedy Menu · Response Matrix · Risk-Based Enforcement Protocol · Sanction Matrix · Least-Restrictive Alternative Screen
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: The mechanism reviews whether the response produced intended safety, fairness, deterrence, repair, or learning without excessive collateral burden, so its operative form is a bounded assessment of existing evidence or work.
Independent corroboration: The frozen evidence defines After-Action Proportionality Review as 'Reviews whether the response produced intended safety, fairness, deterrence, repair, or learning without excessive collateral burden', so its operative form is Assessment, Review & Assurance.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Law & Governance
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Retrospectively testing whether an imposed response was necessary and proportionate to harm and legitimate aims is a canonical legal-governance practice.
Related originating lineages:
- Criminology & Forensic Studies — Deterrence, punishment severity, and unequal application provide a major justice-system use.
- Philosophy — Moral proportionality and balancing of intended benefit against collateral burden supply the normative foundation.
- Psychology — Outcome-bias and hindsight research require preserving the information and reasoning available at decision time before judging the result.
- Public Administration & Policy — Calibration of sanctions and interventions from observed effects contributes administrative review.
Review resolution: Legal proportionality doctrine is primary. Sentencing practice, moral philosophy, hindsight control, and administrative evaluation materially turn the doctrine into a retrospective calibration loop.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
[n1] Outcome bias is the tendency to judge the quality of a decision by how it turned out rather than by the information and reasoning available when it was made. It is the standing hazard of any after-the-fact review, which is why the decision-time frame must be fixed before outcomes are discussed. ↩