{
  "schemaVersion": "1.3",
  "slug": "unique-answer-not-identified",
  "title": "When a Unique Answer Is Not an Identified Answer",
  "shortTitle": "When a unique answer is not an identified answer",
  "url": "https://evidencepress.org/releases/unique-answer-not-identified/",
  "oneLine": "Across APC models, epidemic observation processes, sign-restricted SVARs and aerosol–climate inference, a unique fitted answer can come from a selector rather than from the observations.",
  "abstract": "A fitted model can return one stable number even when the observation regime does not distinguish it from substantively different alternatives. This anonymous, unrefereed synthesis compares four cases: the intrinsic estimator in age–period–cohort models, epidemic severity versus ascertainment, sign-restricted structural vector autoregressions, and aerosol forcing versus equilibrium climate sensitivity. The first three contain exact invariances in the simplified observation maps analysed; the climate case is instead a weak or practical, model-conditional ridge. The paper separates observationally identified content from the coordinates, restrictions, priors, penalties, calibration targets and model-selection rules that narrow an equivalence class or ridge. Its reusable output is a machine-readable Selector Ledger and a four-test publication gate covering invariance, provenance, selector sensitivity and observation design. A purposive 16-record audit illustrates interpretation risks but does not estimate their frequency. The contribution is a comparative bridge and reporting instrument, not a new identification theorem or a claim that all four empirical systems are exactly non-injective.",
  "datePublished": "2026-08-11",
  "dateModified": "2026-08-11",
  "version": "1.0.0-candidate",
  "doi": "10.5281/zenodo.21891811",
  "doiUrl": "https://doi.org/10.5281/zenodo.21891811",
  "conceptDoi": null,
  "pdfUrl": "https://github.com/ipitchford/when-a-unique-answer-is-not-an-identified-answer/releases/download/v1.0.0-candidate/when-a-unique-answer-is-not-an-identified-answer-v1.0.0-candidate.pdf",
  "altPdfUrl": "https://raw.githubusercontent.com/ipitchford/when-a-unique-answer-is-not-an-identified-answer/v1.0.0-candidate/paper/manuscript.pdf",
  "zenodoUrl": "https://zenodo.org/records/21891811",
  "repoUrl": "https://github.com/ipitchford/when-a-unique-answer-is-not-an-identified-answer",
  "releaseUrl": "https://github.com/ipitchford/when-a-unique-answer-is-not-an-identified-answer/releases/tag/v1.0.0-candidate",
  "markdownUrl": "https://evidencepress.org/releases/unique-answer-not-identified/index.md",
  "bibtexUrl": "https://evidencepress.org/releases/unique-answer-not-identified/cite.bib",
  "audioUrl": "https://evidencepress.org/assets/audio/unique-answer-not-identified.mp3?v=f257ffc64a",
  "imageUrl": "https://evidencepress.org/assets/og/unique-answer-not-identified.png",
  "coverArtUrl": "https://evidencepress.org/assets/art/unique-answer-not-identified.svg",
  "media": [
    {
      "type": "audio",
      "url": "https://evidencepress.org/assets/audio/unique-answer-not-identified.mp3",
      "name": "Audio briefing — when a unique answer is not identified",
      "description": "Plain-English OpenAI API synthetic-voice summary of the synthesis, Selector Ledger and assurance boundary; a communication aid, not additional research evidence.",
      "transcriptUrl": "https://evidencepress.org/assets/audio/unique-answer-not-identified.txt"
    },
    {
      "type": "video",
      "url": "https://youtu.be/l_r0fMvFbBQ",
      "name": "Video briefing — The Geometry of Observational Equivalence",
      "description": "Video overview of the identification synthesis and Selector Ledger on the Evidence Press YouTube channel. This is an AI-generated communication summary. It is not additional scientific or mathematical evidence."
    }
  ],
  "authors": [
    "Anonymous author(s)"
  ],
  "license": "CC0-1.0",
  "status": "unrefereed-candidate",
  "verification": {
    "peerReviewed": false,
    "independentlyReproduced": false,
    "formallyVerified": false,
    "internallyReplayed": true,
    "detail": "Anonymous, unrefereed candidate. Producer-side package checks, optimized-interpreter replay, schema validation, negative controls, citation checks and an ex-post quality assessment are recorded. The assessment and tool-assisted reviews occurred inside the producer workflow. No unaffiliated rerun, independent reimplementation, proof-assistant formalization, external field-specialist review, editorial peer review, prevalence study, priority determination or demonstrated research, policy or productivity impact has occurred."
  },
  "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/21891811",
      "note": "The public GitHub prerelease and published Zenodo record expose the same candidate tarball, PDF and SHA256SUMS. Producer-side downloads from both hosts matched the recorded SHA-256 checksums; this establishes public byte availability, not independent scientific reproduction."
    },
    {
      "dimension": "internalReplay",
      "label": "Internal replay",
      "question": "Does the producer’s own pipeline reproduce the stated result from the archived package?",
      "state": "passed",
      "note": "The producer workflow reports ordinary and optimized Python checks, deliberate negative controls, Selector Ledger schema validation, a privacy guard and manifest verification.",
      "evidenceUrl": "https://github.com/ipitchford/when-a-unique-answer-is-not-an-identified-answer"
    },
    {
      "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 running the immutable public candidate package."
    },
    {
      "dimension": "independentReimplementation",
      "label": "Independent reimplementation",
      "question": "Has someone else reached the result from an independent implementation?",
      "state": "not-assessed",
      "note": "The alternate calculations and reviews were produced within one coordinated workflow and are not an independent reconstruction."
    },
    {
      "dimension": "formalVerification",
      "label": "Formal verification",
      "question": "Is a formalised statement machine-checked, and over which trusted base?",
      "state": "not-assessed",
      "note": "No identification statement or Selector Ledger property has been completed in a proof assistant."
    },
    {
      "dimension": "specialistReview",
      "label": "Specialist review",
      "question": "Has a domain specialist assessed the argument?",
      "state": "not-assessed",
      "note": "The ex-post assessment is producer-workflow review; no external APC, infectious-disease, macroeconometric or climate specialist has reviewed the exact candidate."
    },
    {
      "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 comparable venue has run peer 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/when-a-unique-answer-is-not-an-identified-answer/blob/main/README.md",
      "note": "The package uses standard-library Python for the toy checks and supplies commands, manifests and machine receipts, but no independently recreated container, Nix or Guix environment is reported."
    }
  ],
  "provenance": {
    "aiGenerated": true,
    "aiAssisted": true,
    "generatedBy": [
      "AI systems under human direction",
      "OpenAI Codex",
      "Fable, for a terminology suggestion reported by the author(s)"
    ],
    "humanRole": "Research question, source selection, editorial direction, claim-boundary decisions and publication authorization by the human author(s); scholarly attribution remains Anonymous author(s).",
    "disclosure": "The synthesis, toy verifiers, audit, reviews and publication materials were developed in one coordinated AI-assisted producer workflow. Cross-model discussion, hashes, negative controls and public archiving do not constitute independent reproduction, external specialist review or editorial peer review."
  },
  "problem": {
    "name": "Identification and the meanings of an identified answer",
    "url": "https://doi.org/10.1257/jel.20181361"
  },
  "corrections": [],
  "keywords": [
    "identification",
    "identifiability",
    "inverse problems",
    "partial identification",
    "regularization",
    "priors",
    "calibration",
    "age-period-cohort models",
    "epidemic ascertainment",
    "sign-restricted SVARs",
    "climate sensitivity",
    "unrefereed candidate"
  ],
  "keyResults": [
    "A unique computational output does not establish that the observation map is injective; the narrowing may come from a selector such as a coordinate convention, restriction, prior, penalty or calibration rule.",
    "The APC, epidemic-product and orthogonal-rotation examples contain exact invariances in their stated simplified observation maps; aerosol forcing versus climate sensitivity is treated as a weak or practical, model-conditional ridge.",
    "Conditional identification is an orthogonal maintained-assumptions description, not a fifth geometric rung above exact, set, weak or practical identification.",
    "The machine-readable Selector Ledger records identified content, unresolved equivalence or ridge, every narrowing rule, its evidential provenance, required sensitivity checks and rank-restoring observations.",
    "The purposive 16-record audit illustrates five possible interpretation upgrades but supplies no estimate of their prevalence or comparative field quality."
  ],
  "reviews": [],
  "evidencePackage": "An anonymous candidate manuscript and accessible source; a hash manifest for four privately retained frozen source analyses, with the raw exports excluded; a targeted literature matrix and search log; a purposive 16-record audit with a published coding rubric and one high-risk record; a four-case Selector Ledger and JSON Schema; deterministic standard-library Python checks under ordinary and optimized execution; deliberate negative controls; claim, citation, novelty, adversarial and writing reviews; and a complete package manifest. These are producer-side integrity and falsification supports, not independent reproduction of the cited empirical studies.",
  "openProblems": [
    "Obtain one unaffiliated rerun of the exact tagged package and record whether every positive and negative control behaves as declared.",
    "Commission field-specialist readings of the APC, epidemic, SVAR and aerosol–climate sections, with corrections published as linked records rather than silent edits.",
    "Test the Selector Ledger prospectively on analyses not used to design it and compare inter-coder agreement under the published high, medium and low rubric.",
    "Develop formal or executable checks that distinguish likelihood curvature from selector curvature without pretending to validate the substantive observation map automatically.",
    "Evaluate whether differently loaded observations change decisions, not merely parameter precision, in one preregistered applied case."
  ],
  "relatedWorks": [
    {
      "citation": "Lewbel, A. (2019). The Identification Zoo: Meanings of Identification in Econometrics. Journal of Economic Literature, 57(4), 835–903.",
      "url": "https://doi.org/10.1257/jel.20181361"
    },
    {
      "citation": "Maclaren, O. J., & Nicholson, R. (2019). What can be estimated? Identifiability, estimability, causal inference and ill-posed inverse problems.",
      "url": "https://arxiv.org/abs/1904.02826"
    },
    {
      "citation": "Poirier, D. J. (1998). Revising beliefs in nonidentified models. Econometric Theory, 14(4), 483–509.",
      "url": "https://doi.org/10.1017/S0266466698144043"
    },
    {
      "citation": "Tamer, E. (2010). Partial identification in econometrics. Annual Review of Economics, 2, 167–195.",
      "url": "https://doi.org/10.1146/annurev.economics.050708.143401"
    },
    {
      "citation": "Fry, R., & Pagan, A. (2011). Sign Restrictions in Structural Vector Autoregressions: A Critical Review. Journal of Economic Literature, 49(4), 938–960.",
      "url": "https://doi.org/10.1257/jel.49.4.938"
    },
    {
      "citation": "Knutti, R. (2008). Why are climate models reproducing the observed global surface warming so well? Geophysical Research Letters, 35(18), L18704.",
      "url": "https://doi.org/10.1029/2008GL034932"
    },
    {
      "citation": "Giacomini, R., & Kitagawa, T. (2021). Robust Bayesian inference for set-identified models. Econometrica, 89(4), 1519–1556.",
      "url": "https://doi.org/10.3982/ECTA16773"
    }
  ],
  "operatingModel": {
    "version": "1.0",
    "workId": "ep-work:unique-answer-not-identified",
    "attemptIds": [
      "ep-attempt:unique-answer-not-identified"
    ],
    "aims": [
      "science"
    ],
    "artifactRoles": [
      "research-output",
      "evidence-assessment",
      "method-demonstration",
      "communication"
    ],
    "lineageId": null,
    "accelerationPrimitives": [
      "identification-gate",
      "partial-identification",
      "assurance-vector",
      "agent-readable-research-object"
    ],
    "decisionObject": {
      "type": "reusable-method",
      "description": "A Selector Ledger and four-test publication gate for separating observation-supported content from selector-supported narrowing.",
      "scope": "Applied inverse problems with an explicit estimand, observation regime and defensible class of selectors; demonstrated here through four heterogeneous cases."
    },
    "bottleneckTargeted": [
      "assurance",
      "publication"
    ],
    "semanticBridge": {
      "state": "explicit",
      "description": "The common forward-map language is applied exactly only to the stated APC, epidemic-product and rotation maps. The aerosol–climate case is marked as a weak or practical ridge conditional on its model and observation regime.",
      "remainingRisks": [
        "The cross-field abstraction may omit domain-specific distinctions that change the interpretation of a selector.",
        "The purposive audit and its risk codes require external field review and independent coding before any prevalence or comparative-quality claim.",
        "Machine-readable provenance cannot determine whether a scientific restriction or prior is substantively true."
      ]
    },
    "humanJudgmentGates": [
      "Assess whether each field-specific observation map and estimand have been represented faithfully.",
      "Judge whether a selector is scientifically warranted and whether the reported claim preserves its conditionality.",
      "Evaluate novelty and priority beyond the bounded public-corpus search.",
      "Authorize scholarly attribution, public release and any later applied recommendation."
    ],
    "parentLinks": [],
    "assuranceTarget": {
      "dimensions": [
        "independentRerun",
        "specialistReview",
        "semanticValidation",
        "noveltyAssessment",
        "priorityAssessment"
      ],
      "nextAction": "Run the exact tagged package unaffiliated, then obtain field-specific reviews of the four case mappings and a second independent coding of the audit rubric.",
      "claimCeiling": "A comparative bridge and reporting instrument with producer-side checks; not a new identification theorem, prevalence estimate, priority certificate or demonstrated improvement in research practice."
    },
    "impactClaims": [
      {
        "id": "science-no-impact",
        "aim": "science",
        "outcome": "Improved claim–evidence alignment and reduced selector-to-data upgrading",
        "setting": "Cross-field applied research using inverse models, regularisation, priors, restrictions or calibration",
        "status": "NO_IMPACT_EVIDENCE",
        "designClass": "none",
        "comparator": "No matched synthesis, reporting workflow or field-practice comparator was evaluated.",
        "estimand": "No effect on error rates, review time, correction rates or downstream decisions was estimated.",
        "evidenceRefs": [],
        "registeredDesignRef": null,
        "independentAssessment": null
      }
    ]
  }
}