{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp05_complete_proposal_portfolio20_20260803","cell_id":"layer_decay_and_expiration_management__psychology","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_id":"psychology_psychometric_norm_lifecycle_registry_v0","proposal_index":3,"version":0,"title":"Psychometric Norm Lifecycle Registry and Scoring Gate","problem":"Psychological tests can acquire successive norm tables, scoring transformations, cutoff rules, and population-specific calibration files as instruments and reference populations change. Older norm layers may remain available in manuals, spreadsheets, analysis scripts, or scoring systems without an explicit current, superseded, restricted, or archived status. A practitioner or researcher can therefore apply a discoverable but context-mismatched norm as if it were current. Deleting old norm layers is also unsafe because prior reports, longitudinal comparisons, audits, and study reproductions may depend on the exact historical scoring procedure.","actors":["Psychometrician responsible for a test or scale","Test publisher or assessment-program owner","Psychologist or researcher selecting a scoring norm","Software administrator maintaining scoring tools","Records steward responsible for reproducibility and authorized retention"],"observable_state":"For a given psychological measure, multiple norm and scoring packages coexist without a complete identity map. Some lack a collection period, instrument-version binding, intended population, last validation date, supersession link, or explicit lifecycle state. Scoring outputs may record the test name but not the exact norm-package identifier. Observable indicators include duplicate files with unclear authority, use of older packages in new scoring runs, discrepancies between software and manual tables, broken links from historical reports to their scoring basis, and norm layers that remain selectable after their supported context changes.","consequence":"Current assessments can be interpreted against an unintended reference layer, separate users can produce inconsistent standardized results from the same responses, and historical outputs can become difficult to reconstruct. Unqualified deletion can instead break audits, longitudinal interpretation, or reproduction of earlier research.","affected_objective":"Ensure that current scoring uses an explicitly authorized and context-matched norm layer while preserving exact historical scoring dependencies for reconstruction and reproducibility.","intervention":"Establish a Psychometric Norm Lifecycle Registry connected to a scoring gate. Each norm package receives an immutable identifier and records its instrument version, reference sample definition, data-collection period, intended uses and populations, transformations, cutoffs, validation evidence, software dependencies, last review, successor, and lifecycle state: active, restricted, review-due, superseded, quarantined, or archived. A class-specific review lease moves a package to review-due when its validation window ends or its bound instrument changes; it does not declare the norms invalid. Before a new scoring run, the gate requires an active package whose declared context matches the requested instrument and use, or a documented expert override. Superseded packages disappear from ordinary selection but remain addressable by identifier for authorized reconstruction. Dependency checks block removal while reports, studies, longitudinal series, or audit procedures still require the package. Quarantine, disposition markers, tiered archives, and restore drills preserve reversibility and historical interpretation.","structural_mapping":[{"archetype_element":"Sequential deposits","domain_realization":"Successive norm tables, scoring transformations, cutoffs, and calibration packages are added as instruments, samples, and intended uses change."},{"archetype_element":"Changed usefulness or validity","domain_realization":"A norm package's current standing can change when its instrument version, reference population, validation context, or supported scoring software changes."},{"archetype_element":"Stale layers masquerading as current","domain_realization":"Older scoring files remain searchable or selectable without a visible supersession or restriction state."},{"archetype_element":"Layer inventory and identity map","domain_realization":"The registry assigns every norm package an immutable identifier and resolves copies across manuals, files, scripts, and scoring systems."},{"archetype_element":"Age and expiration rule","domain_realization":"A review lease and context-change triggers move packages to review-due rather than using age as an automatic invalidation or destruction rule."},{"archetype_element":"Dependency and reconstruction check","domain_realization":"Reports, publications, longitudinal series, software, and audits are checked for references before a package can leave recoverable storage."},{"archetype_element":"Differentiated disposition","domain_realization":"A package can remain active, become restricted, be superseded, enter quarantine, or move to archive according to its present use and dependencies."},{"archetype_element":"Preservation exception","domain_realization":"A package needed to reproduce a published analysis, explain a prior report, or maintain an authorized longitudinal series can be retained by a named, reviewable exception."},{"archetype_element":"Deletion evidence and rollback","domain_realization":"Disposition markers, reversible quarantine, and archived package restoration preserve lineage after a layer leaves ordinary scoring use."}],"mechanism_mapping":[{"mechanism_slug":"stale_layer_detection_dashboard","role":"Maintains the identity-resolved inventory and flags packages with expired reviews, instrument mismatches, missing provenance, contradicted metadata, broken dependencies, or unrecorded successors.","counterfactual_removal":"Without the inventory and staleness surface, obsolete or context-mismatched scoring layers remain distributed across tools and cannot be systematically reviewed."},{"mechanism_slug":"time_to_live_ttl_policy","role":"Assigns each norm class a review lease; expiration changes the package to review-due and blocks default prospective selection until authorized review, renewal, restriction, or supersession.","counterfactual_removal":"Without a preassigned review trigger, a once-authorized norm package can remain prospectively active indefinitely despite changes in its instrument or supported context."},{"mechanism_slug":"retention_schedule","role":"Maps package classes to active-review intervals, minimum reconstruction periods, archival requirements, and named preservation exceptions for studies, reports, audits, or longitudinal series.","counterfactual_removal":"Without a governing matrix, different custodians could apply inconsistent keep, archive, and destruction rules to equivalent scoring dependencies."},{"mechanism_slug":"dependency_safe_delete_check","role":"Traces inbound references from reports, scoring scripts, publications, longitudinal datasets, and audit procedures and blocks irreversible removal while any authorized dependency remains.","counterfactual_removal":"Without the inbound-reference gate, cleanup could make a prior psychological result impossible to reproduce or explain."},{"mechanism_slug":"lifecycle_storage_tiering_policy","role":"Keeps active norm packages in the scoring tier, moves restricted and superseded packages to controlled archival tiers, and retains exact transformations and documentation together.","counterfactual_removal":"Without tiering, preserving historical packages would either clutter ordinary scoring selection or create pressure to delete reconstruction material."},{"mechanism_slug":"soft_delete_quarantine_window","role":"Places an approved package removal into reversible quarantine, with the grace period based on the number and importance of unresolved or recently retired dependencies.","counterfactual_removal":"Without quarantine, an incorrect dependency judgment could irreversibly remove a scoring basis before the omission is discovered."},{"mechanism_slug":"tombstone_or_deletion_marker","role":"Leaves a durable record at the package identifier stating its disposition, authorizer, reason, date, archive location, and successor package.","counterfactual_removal":"Without a marker, a failed lookup could not distinguish a deliberately superseded package from a package that never existed, and old references would lack a resolution path."},{"mechanism_slug":"archive_restore_test","role":"Samples archived norm packages across ages and formats and executes retrieval, parsing, scoring of a fixed test vector, and reconnection to its documentation and dependency records.","counterfactual_removal":"Without an end-to-end restore exercise, archival metadata could remain intact while the actual scoring transformation becomes unreadable or non-executable."}],"causal_chain":["Psychological measurement programs deposit successive norm and scoring layers over time.","The registry gives each layer an immutable identity, declared context, provenance, dependencies, review lease, successor relationship, and lifecycle state.","Context changes and lease expiration surface packages for review instead of allowing past authorization to persist silently.","The scoring gate excludes review-due, superseded, or context-mismatched packages from default prospective use while permitting documented expert review.","Authorized psychometric reviewers renew, restrict, supersede, quarantine, or archive each package after inspecting its evidence and inbound dependencies.","Ordinary scoring selection presents fewer ambiguous or stale layers, and each resulting output records the exact package identifier used.","Historical packages remain recoverable by identifier for authorized reproduction, audit, and longitudinal interpretation.","Disposition markers and restore drills preserve the lineage and usability of layers removed from active scoring."],"baseline":"Norm tables and scoring rules are distributed across test manuals, local spreadsheets, statistical scripts, and vendor software. Version labels may identify editions, but current authority, intended context, review status, dependencies, and successor relationships are not necessarily governed as one lifecycle. Software updates may replace a default package while historical reports retain only a test name or score, and manual cleanup depends on individual custodians.","nearest_rivals":["Ordinary source-code version control that records file revisions but does not authorize norm contexts or gate prospective scoring","Periodic test renorming that produces a new reference layer without governing the disposition and dependencies of every preceding layer","Test manuals and technical documentation that describe available norms but do not enforce lifecycle states across local scoring copies","Automatic scoring-software updates that change the default package without preserving an inspectable dependency and reconstruction path","General records-retention policies that govern files by document class but do not distinguish active scoring authority from historical reproducibility"],"remaining_contrastive_claim":"The proposal's testable contrast is lifecycle control at the level of executable psychometric norm packages: identity-resolved inventory, expiring prospective authority, context-gated selection, dependency-constrained retirement, and tested historical restoration form one causal path. Versioning, renorming, documentation, software updating, or generic retention alone does not provide that combination.","authority_safety":{"decision_authority":"A designated psychometric authority approves active, restricted, or superseded status and the contexts in which a package may be used. Practitioners and researchers may select only authorized packages or submit a documented override for expert review. Records stewards may execute approved tiering or quarantine but cannot decide psychometric validity. The registry cannot independently infer that a norm is valid or invalid.","authorized_first_step":"Construct a read-only inventory for one synthetic or appropriately authorized test family, assign provisional identifiers and lifecycle metadata, and evaluate a mock scoring gate without changing any production scoring system, report, research result, or source file.","excluded_actions":["Automatically declaring a population norm valid or invalid","Recalculating or revising historical psychological reports without separate authorization","Deleting raw assessment data, norm samples, signed reports, or published analysis artifacts","Treating chronological age as sufficient evidence for psychometric obsolescence","Allowing an automated score to override professional interpretation","Blocking emergency access to information required under an existing authorized procedure","Using archived packages prospectively without explicit context review","Changing test cutoffs or transformations through metadata cleanup alone"],"halt_rollback":"Halt if the registry assigns ambiguous identities, hides the package required to reconstruct a prior result, permits a context-mismatched package as active, or fails to restore a quarantined or archived scoring transformation. Disable the mock gate, return all packages to their prior visibility, preserve the audit output, and revert to the existing scoring-selection procedure."},"negative_tests":{"strongest_counterevidence":"Existing edition controls, manuals, and professional judgment may already make package selection unambiguous, while an additional gate could create false confidence that contextual validity has been resolved merely because metadata fields are complete.","problem_falsifier":"A bounded audit finds that every scoring output already records an exact immutable norm identifier, only context-authorized packages are prospectively selectable, superseded packages are visibly distinguished, and all sampled historical results can be reconstructed without ambiguity or missing dependencies.","intervention_falsifier":"In a blinded scoring-selection and reconstruction task, the registry and gate do not reduce context-mismatched package selections or improve historical reconstruction, or they increase valid-use blocks, expert override burden, identity errors, scoring discrepancies, or restoration failures beyond predeclared tolerances.","risks":["Lifecycle status could be mistaken for a guarantee of psychometric validity.","Incomplete context metadata could falsely authorize a mismatched norm package.","A scoring gate could delay a legitimate assessment use not represented in the registry.","Overbroad preservation exceptions could retain sensitive norm or assessment data unnecessarily.","An undiscovered local script or printed table could bypass the registry.","Dependency mapping may miss reports that recorded scores without package identifiers.","Archival formats or software environments may become non-executable despite successful metadata checks.","Superseding a package could complicate longitudinal comparisons that intentionally require a stable historical norm.","Registry access could reveal restricted test materials or sensitive reference-sample information."]},"next_evidence_step":"Using synthetic response data and one synthetic or appropriately authorized family of historical norm packages, create scenarios that vary instrument edition, intended population, assessment date, longitudinal purpose, supersession state, and reconstruction dependency. Randomly assign qualified evaluators to the existing file-and-manual presentation or a read-only registry mock-up. Measure selection of the context-authorized package, inappropriate use of superseded packages, time to identify the scoring basis, valid-use blocks, agreement about required preservation, and exact reconstruction of fixed historical scores. Perform an end-to-end restore of one copied archived package and verify its output against a predeclared test vector. Do not change live scoring, reports, or source records.","prior_art_status":"UNSEARCHED","diversity_from_prior_proposals":"Proposal 1 manages clinician-authored patient formulations so outdated hypotheses cease competing with current clinical reasoning while signed history remains reconstructable. Proposal 3 instead manages shared psychometric norm and scoring packages as methodological infrastructure; its central control is a context-sensitive scoring gate operated under psychometric authority, not lifecycle review of patient claims. Proposal 2 manages participant-owned behavior-change commitments under a bounded active-attention budget; proposal 3 has no goal-slot budget, cue-management objective, participant-controlled sunset decision, or commitment portfolio. Its causal path runs from accumulated calibration layers through expiring scoring authority and context gating to consistent prospective scoring, with dependency-preserving archives supporting reproducibility. It is independently adoptable by a test owner or research program without either a clinical formulation system or a behavior-change ledger.","revision_record":{"parent_version":null,"progress_targets_addressed":["Generate proposal index 3 as one complete candidate","Address a psychology problem materially different from proposals 1 and 2","Provide a distinct intervention and causal path faithful to layer lifecycle management","Explain diversity from each earlier sealed proposal","Specify operational authority, safeguards, falsifiers, rivals, and bounded evidence"],"conceptual_changes":[],"operational_changes":[],"evidence_changes":[],"claim_changes":[]}}