{"actors":["Administrative agency enforcement analysts who run cross-program compliance screens","Case-intake supervisors who authorize investigations","Agency counsel who assess proposed coercive action","Independent confirmation officers who conduct targeted rechecks","Regulated entities whose licenses, funds, or reputations may be affected","Administrative review officials who audit the evidentiary record"],"affected_objective":"Calibrate the evidentiary status of data-selected compliance allegations before they support subpoenas, sanctions, license restrictions, or public accusations.","arm":"COMMON_P1","authority_safety":{"authorized_first_step":"The agency's analytics-governance lead may conduct a retrospective, read-only review of a bounded set of closed screening-to-enforcement files and prototype a shadow registry without altering any case status.","decision_authority":"Only the official already empowered by applicable law and agency procedure may authorize an investigation or coercive action; the registry records evidentiary status but neither creates jurisdiction nor makes the legal decision.","excluded_actions":["Automatically closing or declining a matter because it fails a multiplicity rule","Using an exploratory flag as the sole basis for compulsory process, sanction, license restriction, funding withdrawal, or public naming","Reopening closed matters during the first evidence step","Changing statutory burdens, appeal rights, disclosure duties, or limitation periods","Exposing privileged, confidential, or protected personal information in the registry"],"halt_rollback":"Halt the pilot if it exposes protected information, delays a legally mandatory response, misclassifies independently sourced evidence as screen-selected, or is treated as an automated merits decision; disable the shadow workflow, preserve the audit log under existing retention rules, and return to ordinary authorized review."},"baseline":"An analyst searches a compliance warehouse across programs, indicators, thresholds, jurisdictions, entity types, and time windows, then forwards the strongest anomaly. The referral records the selected result and supporting rationale but not necessarily the full opportunity set, and supervisors may treat the threshold crossing as if it were a single pre-specified test.","candidate_id":"multiple_testing_discipline__law_governance__COMMON_P1","causal_chain":["Agency analysts inspect many related compliance indicators, entity segments, thresholds, jurisdictions, and time windows.","Each additional look creates another opportunity for an innocent entity or program to exhibit an apparently exceptional pattern by chance or data artifact.","When unselected and null looks are absent from the referral, the selected anomaly appears more exceptional than its search history warrants.","A supervisor may consequently grant the anomaly decision-ready status and authorize costly investigation or coercive escalation.","A screening-family docket makes the complete family and analytic choices visible, applies the declared error-risk rule, and labels selected anomalies as exploratory or adjusted discoveries.","An independently assigned officer then tests a frozen allegation against fresh-period records or independently obtained case evidence.","Only an allegation satisfying the predeclared confirmation gate may be presented as screen-confirmed evidence for an authorized official's substantive legal judgment."],"cell_id":"multiple_testing_discipline__law_governance","consequence":"A chance pattern, coding artifact, or selectively favorable slice can acquire the appearance of a uniquely strong compliance signal and contribute to investigation costs, reputational injury, unequal scrutiny, or coercive government action without an evidentiary burden matched to the search that produced it.","diversity_from_prior_proposals":"Comparison with prior proposals was not performed under runtime isolation; this candidate is specifically instantiated as a pre-coercion evidentiary-status gate for administrative compliance screens, not as a general court filing rule or a rule for evaluating one pre-specified allegation.","experiment_id":"eoa_inverse_innovation_exp13_second_slot_policy60_20260806","intervention":"Create a screening-family docket for data-generated administrative enforcement referrals. Before reviewing results, the screening owner defines the related indicators, populations, thresholds, jurisdictions, models, and time windows; every attempted look and material analytic change is appended to an access-controlled record. The agency counsel responsible for evidence governance selects a declared familywise, discovery-rate, or exploratory-only policy according to the contemplated downstream action. The system reports both the selected anomaly and its family-level status, prohibits an exploratory label from being represented as confirmed, and freezes the allegation and confirmation protocol. Before coercive escalation based materially on the screen, an officer who did not select the anomaly tests it using a fresh time period or independently obtained case evidence. The complete discovery and confirmation history accompanies the referral to the legally authorized decision-maker.","mechanism_mapping":[{"counterfactual_removal":"Without the registry, null screens, abandoned thresholds, and repeated slices can disappear, leaving reviewers unable to reconstruct how many opportunities produced the referral.","mechanism_slug":"claim_registry","role":"Maintains an access-controlled docket of planned and attempted screens, analytic changes, selected allegations, owners, and evidentiary statuses."},{"counterfactual_removal":"Without a multiplicity adjustment, a threshold suitable for one compliance comparison remains unchanged even when the agency creates many opportunities for a false flag.","mechanism_slug":"bonferroni_like_correction","role":"Provides a strict family-level screen when even one false allegation could lead to severe or difficult-to-reverse government action."},{"counterfactual_removal":"Without fresh evidence, the same records used to discover the anomaly also appear to confirm it, preserving selection-driven overstatement.","mechanism_slug":"holdout_validation","role":"Tests the frozen allegation against a later, previously unused reporting period or an evidence partition inaccessible during screening."},{"counterfactual_removal":"Without a targeted second stage, broad screening output can move directly from lead generation to case escalation.","mechanism_slug":"confirmatory_follow_up","role":"Requires an independently assigned officer to examine the frozen allegation under predeclared criteria before screen-confirmed status is available."}],"nearest_rivals":["A single-claim evidentiary threshold for one allegation does not address the many unreported opportunities that generated a selected allegation.","A reproducible analytics protocol can recreate every screen while still allowing the most favorable result to be treated as uniquely probative.","Confounder review asks whether an observed association has an alternative cause, whereas this intervention first asks how many chances the agency had to find an impressive association.","Ordinary supervisory review assesses legal sufficiency and investigative value but may not reconstruct or discipline the full search family.","A general reason-giving requirement explains why officials acted but does not by itself preserve null screens or separate discovery evidence from confirmation evidence."],"negative_tests":{"intervention_falsifier":"In a shadow comparison, the docket fails if reviewers given the docket cannot reliably distinguish exploratory, multiplicity-adjusted, and independently confirmed referrals, or if undocumented analytic looks remain common enough that the represented claim family materially understates the actual search.","problem_falsifier":"The proposed problem is absent in the reviewed pathway if consequential referrals consistently arise from one genuinely pre-specified indicator and population, or from independently sourced complaints and case evidence, with no material selection across alternative outcomes, segments, models, thresholds, or time windows.","risks":["Analysts may evade the rule by splitting one search into artificially narrow claim families.","A strict rule may suppress useful leads that could be followed up cheaply.","Fresh-period data may differ for substantive reasons, making failed confirmation ambiguous.","Confirmation officers may be influenced by the original referral and cease to be meaningfully independent.","The registry may expose privileged investigative strategy or protected entity information if access controls are inadequate.","Officials may treat an adjusted screen as proof of legal liability despite unresolved measurement, confounding, materiality, or due-process questions.","Administrative burden may encourage off-system analyses that worsen the discovery record."] ,"strongest_counterevidence":"The strongest counterevidence would be a bounded file audit showing that enforcement referrals already disclose every material screen and analytic choice, use a policy matched to the full family, preserve null results, and require genuinely independent fresh evidence before any screen materially supports coercive action."},"next_evidence_step":"Select, under existing audit authority, 20 closed referrals from one compliance-screening program over a fixed six-month period. For each, reconstruct from retained query logs and case files the indicators, segments, thresholds, models, jurisdictions, and time windows inspected; compare that inventory with the referral record; and have two authorized reviewers independently assign exploratory, adjusted-discovery, or confirmed status using a paper prototype. Record agreement, missing-look counts, review time, confidentiality failures, and whether the proposed gate would have conflicted with a mandatory deadline. Make no change to any case or enforcement outcome.","observable_state":"For each screened referral, an authorized reviewer can inspect a timestamped docket containing the declared claim family, all executed and materially considered screens, threshold and model changes, null and selected results, applicable error-risk policy, adjustment or exploratory-only treatment, current status label, frozen confirmation question, source and separation of confirmation evidence, reviewer identities, and the authorized official's ultimate decision rationale.","prior_art_status":"UNSEARCHED","problem":"An administrative agency searches a shared compliance dataset across many indicators, regulated-entity categories, jurisdictions, thresholds, models, and time windows, then elevates the most exceptional-looking anomaly as an enforcement lead. Because the referral displays the chosen anomaly without the full family of failed or unselected looks, supervisors can interpret a discovery produced by broad search as though it were a uniquely planned allegation and use it to support consequential process.","proposal_index":1,"remaining_contrastive_claim":"The candidate adds value only where an administrative allegation is selected from a family of data searches and may contribute to coercive action: its distinctive requirement is that the whole search family determine the allegation's evidentiary status and that fresh, separated evidence precede screen-confirmed status. It does not determine legal liability, cure biased data, establish practical materiality, or replace the authorized official's judgment.","revision_record":{"claim_changes":["Initial version; no claims of novelty, prevalence, demand, or effect size are made."],"conceptual_changes":["Initial version maps multiplicity discipline to the transition from administrative compliance screening to coercive enforcement."],"evidence_changes":["Prior art remains unsearched; the proposal specifies a bounded retrospective file audit as the first evidence step."],"operational_changes":["Initial version limits implementation to a read-only shadow registry and paper status review of closed files."],"parent_version":null,"progress_targets_addressed":["Concrete many-look administrative problem","Observable discovery and confirmation record","Authority-preserving pre-coercion safeguard","Separate problem and intervention falsifiers","Bounded non-interventional evidence step"]},"schema_version":1,"structural_mapping":[{"archetype_element":"Define the claim family","domain_realization":"Before results are examined, define the compliance indicators, regulated populations, jurisdictions, models, thresholds, and time windows treated as one screening family."},{"archetype_element":"Inventory attempted looks","domain_realization":"Append executed queries, filters, alternate specifications, threshold changes, null screens, and abandoned slices to the referral's screening-family docket."},{"archetype_element":"Set the error-risk policy","domain_realization":"Tie the policy to the contemplated governmental consequence: strict family-level prevention for severe coercive action, discovery-rate treatment for lead triage, or explicit exploratory status where inexpensive follow-up remains available."},{"archetype_element":"Apply a multiplicity-aware rule","domain_realization":"Evaluate the selected compliance anomaly under the declared family-level adjustment or withhold confirmatory standing pending a fresh-evidence stage."},{"archetype_element":"Label claim status","domain_realization":"Mark each allegation as exploratory, multiplicity-adjusted discovery, independently confirmed, rejected, or superseded in the case-intake record."},{"archetype_element":"Confirm before costly action","domain_realization":"Freeze the selected allegation and have a separately assigned officer test it against an unused reporting period or independently obtained case evidence before it can receive screen-confirmed status."},{"archetype_element":"Preserve the discovery record","domain_realization":"Retain the docket, status changes, null results, confirmation materials, access history, and final authorized rationale under existing records, privilege, and confidentiality rules."}],"title":"Screening-Family Docket for Data-Selected Administrative Enforcement Leads","version":0}