{
  "schemaVersion": "1.6",
  "slug": "wales-20mph-attribution-manuscript",
  "title": "Wales 20 mph Casualty Attribution Audit: Companion Manuscript",
  "shortTitle": "Wales 20 mph audit: the manuscript",
  "url": "https://evidencepress.org/releases/wales-20mph-attribution-manuscript/",
  "oneLine": "The paper behind the August 2026 Wales 20 mph audit: the 28% first-year fall reproduces as a description, but the published totals identify neither the size nor the sign of the policy's effect on the roads that changed.",
  "abstract": "Companion manuscript to the Evidence Press evidence package of 11 August 2026 (DOI 10.5281/zenodo.21894035), which was released without a paper. On 17 September 2023 the default speed limit on restricted roads in Wales fell from 30 mph to 20 mph, and a provisional Welsh Government release reported that casualties on roads recorded as 20 or 30 mph were 28% lower in 2023 Q4 to 2024 Q3 than in the preceding four quarters. Using the package's corpus of 25 registered sources frozen on 11 August 2026, the manuscript reproduces the headline as a description (2,402 to 1,725 casualties, exactly -33850/1201 per cent) and reports that later frozen vintages give about -26.10% and -24.18%. It separates that statistic from the registered causal estimand, the effect on police-recorded casualties on road segments whose limit actually changed, and gives a constructive witness: three hypothetical worlds reproduce the observed aggregates yet imply effects of -50%, 0% and +100%. Within the frozen evidence and admitted rival set neither the effect nor its sign is identified, and no empirical bound is registered. The manuscript adds no new empirical analysis and does not change the package. It does not show that the policy succeeded or failed.",
  "datePublished": "2026-10-12",
  "dateModified": "2026-10-12",
  "version": "1.0.0-candidate",
  "doi": "10.5281/zenodo.23326288",
  "doiUrl": "https://doi.org/10.5281/zenodo.23326288",
  "conceptDoi": "10.5281/zenodo.23326287",
  "pdfUrl": "https://github.com/ipitchford/wales-20mph-attribution-manuscript/releases/download/v1.0.0-candidate/wales-20mph-attribution-manuscript-v1.0.0-candidate.pdf",
  "altPdfUrl": "https://zenodo.org/records/23326288/files/wales-20mph-attribution-manuscript-v1.0.0-candidate.pdf?download=1",
  "altPdfLabel": null,
  "zenodoUrl": "https://zenodo.org/records/23326288",
  "repoUrl": "https://github.com/ipitchford/wales-20mph-attribution-manuscript",
  "releaseUrl": "https://github.com/ipitchford/wales-20mph-attribution-manuscript/releases/tag/v1.0.0-candidate",
  "markdownUrl": "https://evidencepress.org/releases/wales-20mph-attribution-manuscript/index.md",
  "bibtexUrl": "https://evidencepress.org/releases/wales-20mph-attribution-manuscript/cite.bib",
  "audioUrl": "https://evidencepress.org/assets/audio/wales-20mph-attribution-manuscript.mp3",
  "imageUrl": "https://evidencepress.org/assets/og/wales-20mph-attribution-manuscript.png?v=1f1efb2f88",
  "coverArtUrl": "https://evidencepress.org/assets/art/wales-20mph-attribution-manuscript.svg?v=e7c3856a4d",
  "media": [
    {
      "type": "audio",
      "url": "https://evidencepress.org/assets/audio/wales-20mph-attribution-manuscript.mp3",
      "name": "Audio briefing — Wales 20 mph casualty attribution audit: companion manuscript",
      "description": "Synthetic plain-English explanation, not additional evidence.",
      "transcriptUrl": "https://evidencepress.org/assets/audio/wales-20mph-attribution-manuscript.txt"
    }
  ],
  "authors": [
    "Anonymous"
  ],
  "license": "CC0-1.0",
  "status": "unrefereed-candidate",
  "verification": {
    "peerReviewed": false,
    "independentlyReproduced": false,
    "formallyVerified": false,
    "internallyReplayed": true,
    "detail": "Unrefereed candidate. Companion manuscript reporting the frozen records of the 11 August 2026 evidence package without new empirical analysis. Producer-side replay passes; the review on record is internal AI-assisted adversarial review. No independent rerun, independent reimplementation, formal verification, external specialist review or editorial peer review has been established. The positive 28% attribution remains unsupported; no policy effect is estimated."
  },
  "assurance": [
    {
      "dimension": "availability",
      "label": "Availability and archiving",
      "question": "Is the evidence package publicly retrievable from an archive under a persistent identifier?",
      "state": "passed",
      "evidenceUrl": "https://zenodo.org/records/23326288",
      "note": "Manuscript PDF, tagged archive and checksums published on GitHub and Zenodo; public downloads matched the local SHA-256 ledger. Availability is not correctness or peer review."
    },
    {
      "dimension": "internalReplay",
      "label": "Internal replay",
      "question": "Does the producer’s own pipeline reproduce the stated result from the archived package?",
      "state": "passed",
      "note": "The evidence package's validation, citation, replay, table, ordinary and optimized test, optimized-control and lifecycle checks were rerun by the producer on 12 October 2026 from a clean clone of its v1.0.0 tag; all passed. Producer-side only.",
      "evidenceUrl": "https://github.com/ipitchford/wales-20mph-attribution-manuscript/tree/v1.0.0-candidate/review/replay-2026-10-12"
    },
    {
      "dimension": "independentRerun",
      "label": "Independent rerun",
      "question": "Has someone else run the supplied implementation and obtained the stated result?",
      "state": "not-assessed",
      "note": "No unaffiliated party has reported rerunning the package."
    },
    {
      "dimension": "independentReimplementation",
      "label": "Independent reimplementation",
      "question": "Has someone else reached the result from an independent implementation?",
      "state": "not-assessed",
      "note": "No unaffiliated implementation of the vintage ladder or witness."
    },
    {
      "dimension": "formalVerification",
      "label": "Formal verification",
      "question": "Is a formalised statement machine-checked, and over which trusted base?",
      "state": "not-assessed",
      "note": "Exact arithmetic and schema guards execute; no proof-assistant theorem."
    },
    {
      "dimension": "specialistReview",
      "label": "Specialist review",
      "question": "Has a domain specialist assessed the argument?",
      "state": "not-assessed",
      "note": "No external human road-safety, official-statistics or causal-inference specialist review; model reviews are internal."
    },
    {
      "dimension": "editorialPeerReview",
      "label": "Editorial peer review",
      "question": "Has a journal or venue run peer review to a decision?",
      "state": "not-assessed",
      "note": "No journal or venue review to a decision."
    },
    {
      "dimension": "dataEnvironmentReproducibility",
      "label": "Data and environment reproducibility",
      "question": "Are data and computational environment pinned well enough to rebuild?",
      "state": "partial",
      "evidenceUrl": "https://github.com/ipitchford/wales-20mph-policy-analysis/blob/v1.0.0/uv.lock",
      "note": "Locked Python environment, frozen source hashes and command vectors; no independently recreated container environment."
    }
  ],
  "provenance": {
    "aiGenerated": true,
    "aiAssisted": true,
    "generatedBy": [
      "OpenAI Codex agent team (evidence package)",
      "Anthropic Claude (manuscript drafting and publication, 12 October 2026)",
      "OpenAI gpt-5.6-sol (adversarial manuscript review)"
    ],
    "humanRole": "Publication authority only; no human research contribution asserted.",
    "reviewStatus": "Unrefereed candidate. Internal AI-assisted review is not external specialist review, independent reproduction or editorial peer review."
  },
  "attributionBoundary": null,
  "problem": null,
  "orderedProductAverage": null,
  "corrections": [],
  "publicCorrections": [],
  "publicCorrectionIndexes": null,
  "pageStructureVersion": "2.0",
  "pageStructureVariant": "policy-identification-case",
  "pageStructureWaivers": [],
  "recordMaturity": "current",
  "metadataProvenance": "Authored 12 October 2026 from the frozen case records of the 11 August 2026 evidence package, the companion manuscript and its review record.",
  "grandfatheredAtSchemaVersion": null,
  "keywords": [
    "Wales",
    "20 mph",
    "road casualties",
    "policy identification",
    "causal inference",
    "non-identification",
    "constructive counterexample",
    "official statistics",
    "companion manuscript"
  ],
  "keyResults": [
    "The provisional first-year comparison reconstructs to 2,402 and 1,725 casualties on roads recorded as 20 or 30 mph, an exact change of -33850/1201 per cent (about -28.18%); later frozen vintages give about -26.10% (StatsWales) and -24.18% (DfT collision rows).",
    "The published statistic covers every road recorded as 20 or 30 mph; the registered estimand is the effect on segments whose limit actually changed, relative to their unobserved no-policy outcome.",
    "Three hypothetical worlds share the observed totals and the same affected/other split yet imply affected-road effects of -50%, 0% and +100%: neither the effect nor its sign is identified, and the values are not estimates or bounds.",
    "Diagnostics (an England comparison whose pre-period differential reverses sign, outcome-scope sensitivity, speed-limit coding corrections) do not restore identification."
  ],
  "reviews": [],
  "evidencePackage": "A 16-page manuscript with source, figure script, an AI-assisted adversarial review record with dispositions, and producer rerun logs of the evidence package's validation, citation, replay, table, test and lifecycle checks on 12 October 2026. The scientific records are those of the evidence package (DOI 10.5281/zenodo.21894035), which this release does not change.",
  "openProblems": [
    "Freeze a versioned historical segment map of roads that changed, kept 30 mph or later reverted, and link collision locations to it.",
    "Specify and independently review a comparison design with several pre-period diagnostics, placebo dates and explicit anticipation and interference assumptions.",
    "Independently reconstruct the source-vintage ladder and reimplement the observation-map witness without producer code."
  ],
  "mathObjects": [],
  "relatedWorks": [
    {
      "citation": "Wales 20 mph Casualty Attribution Audit: A Constructive Non-Identification Result — the evidence package this manuscript accompanies (Evidence Press, 11 August 2026; DOI 10.5281/zenodo.21894035).",
      "url": "https://evidencepress.org/releases/wales-20mph-casualty-attribution/",
      "doi": null
    },
    {
      "citation": "Welsh Government (2025). Police recorded road collisions: July to September 2024 (provisional).",
      "url": "https://www.gov.wales/police-recorded-road-collisions-july-september-2024-provisional-html",
      "doi": null
    },
    {
      "citation": "Office for Statistics Regulation (2026). 20mph speed limit.",
      "url": "https://osr.statisticsauthority.gov.uk/guidance/2026-what-to-look-out-for-20mph/",
      "doi": null
    },
    {
      "citation": "Xu, J. (2026). Replication materials for Vulnerable Road User Safety after Wales's Default 20 mph Reform: Implications for Sustainable Mobility, version 1.1 — a later independent comparison-group study; discussed, not assessed.",
      "url": "https://doi.org/10.5281/zenodo.22259378",
      "doi": "10.5281/zenodo.22259378"
    },
    {
      "citation": "Slower Streets, Safer Streets? APEP 0513 v1 (2026) — earlier Wales/England difference-in-differences evaluation.",
      "url": "https://ape.socialcatalystlab.org/papers/apep_0513/v1/paper.pdf",
      "doi": null
    },
    {
      "citation": "Does the Default Kill? APEP 0734 v1 (2026) — earlier evaluation discussing 20/30 mph reclassification.",
      "url": "https://ape.socialcatalystlab.org/papers/apep_0734/v1/paper.pdf",
      "doi": null
    },
    {
      "citation": "The Speed Penalty. APEP 0744 v1 (2026) — earlier evaluation treating 2023 Q3 as a transition quarter.",
      "url": "https://ape.socialcatalystlab.org/papers/apep_0744/v1/paper.pdf",
      "doi": null
    }
  ],
  "claimAssurance": null,
  "operatingModel": {
    "version": "1.0",
    "workId": "ep-work:wales-20mph-attribution-manuscript",
    "attemptIds": [
      "ep-attempt:wales-20mph-attribution-manuscript-publication"
    ],
    "aims": [
      "policy"
    ],
    "artifactRoles": [
      "research-output",
      "communication"
    ],
    "lineageId": null,
    "accelerationPrimitives": [
      "identification-gate",
      "counterexample-proxy-first",
      "adversarial-controls",
      "assurance-vector",
      "agent-readable-research-object"
    ],
    "decisionObject": {
      "type": "counterexample",
      "description": "Three hypothetical affected-road worlds that share the observed aggregate totals but imply negative, zero and positive effects, set out with the estimand, diagnostics and limitations as a manuscript.",
      "scope": "The first-four-complete-quarters affected-road casualty estimand, frozen official corpus and admitted rival class of the 11 August 2026 evidence package. The worlds are counterexamples, not estimates or bounds."
    },
    "bottleneckTargeted": [
      "assurance",
      "publication"
    ],
    "semanticBridge": {
      "state": "partial",
      "description": "The manuscript separates the combined 20/30 mph descriptive proxy from the affected-segment causal estimand and states the observation map explicitly. The empirical bridge cannot be completed because the frozen corpus lacks a historical treatment map and the affected-road no-policy outcome.",
      "remainingRisks": [
        "Conditional on the registered corpus, observation map and admitted rival class.",
        "No human specialist has assessed the semantic bridge.",
        "The package's round-four review report and review-history field disagree (READY versus REVISION_REQUESTED); the manuscript states this without resolving it."
      ]
    },
    "humanJudgmentGates": [
      "Assess whether the affected-road estimand and rivals represent the policy question adequately.",
      "Evaluate policy merits, which the identification audit does not decide."
    ],
    "parentLinks": [
      {
        "relation": "reuses-evidence",
        "workId": "ep-work:wales-20mph-casualty-attribution",
        "legacyReleaseSlug": null,
        "externalUrl": null,
        "inheritedClaim": "The manuscript reports the parent package's frozen records, witness and claim ceiling without new empirical analysis.",
        "inheritedAssuranceCeiling": "A companion manuscript is not independent confirmation of the parent; both rest on the same producer-side records and replay."
      }
    ],
    "assuranceTarget": {
      "dimensions": [
        "specialistReview",
        "independentReimplementation",
        "independentRerun"
      ],
      "nextAction": "An unaffiliated road-safety statistics or causal-inference specialist should assess the estimand, rival set and witness, and reimplement the vintage ladder without producer code.",
      "claimCeiling": "Within the frozen aggregate evidence and declared rival class, neither the point value nor the sign of the affected-road causal effect is identified. No effect estimate, universal impossibility result, independent reproduction, peer review or policy judgement is established."
    },
    "impactClaims": [
      {
        "id": "wales-20mph-attribution-manuscript-no-impact",
        "aim": "policy",
        "outcome": "Clearer public assessment of causal policy claims",
        "setting": "Companion manuscript to a policy identification audit",
        "status": "NO_IMPACT_EVIDENCE",
        "designClass": "none",
        "comparator": "None.",
        "estimand": "No decision-quality or assessment-time effect estimated.",
        "evidenceRefs": [],
        "registeredDesignRef": null,
        "independentAssessment": null
      }
    ]
  },
  "researchMetrics": {
    "schemaVersion": "1.0",
    "policyUrl": "https://evidencepress.org/api/research-metrics-policy.json",
    "workLedgerUrl": "https://evidencepress.org/api/work-ledger.json",
    "claimCeiling": "These records describe scoped workflow inputs and outcomes. They support calibration and reference-class construction, but do not by themselves establish research acceleration, theorem correctness, novelty, independent assurance or impact.",
    "attempts": [
      {
        "attemptId": "ep-attempt:wales-20mph-attribution-manuscript-publication",
        "status": "result-ready",
        "resultClass": "positive",
        "metrics": {
          "schemaVersion": "1.0",
          "measurementScope": "assurance-through-publication",
          "scopeBoundary": "Site-ledger registration onward; earlier preparation is left-censored.",
          "forecast": {
            "frozenAt": "2026-10-12T21:06:11.000Z",
            "procedureClass": "full-candidate",
            "targetOutcome": "Immutable companion-manuscript publication and canonical site readback",
            "expectedActiveMinutes": 90,
            "plausibleLowMinutes": 60,
            "plausibleHighMinutes": 150,
            "expectedUnattendedWaitMinutes": 30,
            "referenceClass": {
              "label": "Remaining full-candidate publication",
              "sampleSize": 0,
              "basis": "Procedural forecast, not measured benchmark"
            },
            "fermiComponents": [
              {
                "component": "Provider publication and asset readback",
                "count": 1,
                "lowMinutesPerUnit": 10,
                "centralMinutesPerUnit": 20,
                "highMinutesPerUnit": 35,
                "basis": "Small archive; provider latency remains variable."
              },
              {
                "component": "Media, page and browser QA",
                "count": 1,
                "lowMinutesPerUnit": 20,
                "centralMinutesPerUnit": 30,
                "highMinutesPerUnit": 50,
                "basis": "New release, banner, thumbnail and audio."
              },
              {
                "component": "Both guarded A/B and C/D deployment cycles and closeout",
                "count": 1,
                "lowMinutesPerUnit": 30,
                "centralMinutesPerUnit": 40,
                "highMinutesPerUnit": 65,
                "basis": "Required CI, preservation and canonical readback."
              }
            ],
            "probabilityPositiveSignal": 0.9,
            "probabilityTargetClosure": 0.8,
            "probabilityHorizonMinutes": 480,
            "assumptions": [
              "The frozen manuscript remains unchanged and provider, CI and deployment routes remain accessible."
            ],
            "reforecastTriggers": [
              "A substantive integrity finding, provider failure or CI failure."
            ],
            "stopRule": "No hard cap; complete mandatory gates or report a precise blocker."
          },
          "outcome": null
        },
        "missingnessReason": null
      }
    ]
  }
}