Research study register

Register version: 0.1.0 Evidence cutoff: 2026-08-31 Status: public claims-control record; no interview study is substantiated or published

The register records commissioned, internal, and independent studies that may be cited by the benchmark. It is not a list of marketing claims. A study appears here when a claim is received, before evidence is accepted, so missing evidence remains visible.

The machine-readable authority for the current entry is the held synthetic-safe intake fixture at ../../FIXTURES/research-study/valid/study-register.unverified.json. The companion evidence-package manifest records each absent artifact class rather than treating missing materials as zero-valued evidence. Real participant records must never be committed here.

Registered study FDE-JA-001

Field	Current record
Permanent title	Commissioned AI practitioner and executive job-analysis interviews
Sponsor / commissioner	LockedIn Labs, requestor-reported; documentary confirmation pending
Collector	Unknown in repository; roster and contract requested
Intended use	Inform the FDE role boundary, task/work-product inventory, capability ontology, and later assessment content blueprint
Design timing	Retrospective reconstruction; no claim of preregistration
Requested practitioner claim	At least 500 AI practitioners interviewed
Requested executive claim	At least 50 executives interviewed
Distinct people required for both claims	At least 550 under the mutually exclusive primary-cohort rule
Participant records received	0 at cutoff
Records eligible for public count	0 at cutoff
Current state	`claim_received`
Publication state	`blocked`
Public finding	None
Score or hiring authority	None

“0 records received” describes repository evidence at the cutoff; it does not assert that no interviews occurred. The original work may exist outside this repository. Its existence, population, conduct, and findings remain unverified until intake.

Claim ledger

Claim ID	Requestor-reported statement	Evidence state	Minimum needed	Current disposition
FDE-JA-001-C01	“LockedIn Labs interviewed at least 500 AI practitioners.”	Unverified requestor assertion	500 distinct complete, consent-eligible, practitioner-eligible records plus frozen count receipt	Blocked from publication
FDE-JA-001-C02	“LockedIn Labs interviewed at least 50 executives.”	Unverified requestor assertion	50 distinct complete, consent-eligible, executive-eligible records plus frozen count receipt	Blocked from publication
FDE-JA-001-C03	“The LockedIn FDE Benchmark is based on this research.”	Not evaluated	C01/C02 as applicable, analyzable instruments/records, coding and agreement, content traceability, independent review, governance release	Blocked from publication
FDE-JA-001-C04	“The benchmark accurately predicts hiring or job performance.”	Unsupported by this design	Separate intended-use criterion/predictive validation and consequence evidence	Prohibited under this study

Intake blockers

Commission/scope, collector identity, dates, funding, and caller roster are absent.
Sampling frame, recruitment messages, contact disposition log, incentives, inclusion/exclusion rules, and nonresponse are absent.
Consent notice, versions, recording permissions, withdrawal/correction path, and applicability determination are absent.
Interview guides, translations, version-to-session mapping, interviewer training, and deviations are absent.
No deidentified interview records, participant deduplication keys, role-eligibility evidence, session metadata, artifact digests, or controlled data manifest have been received.
No codebook, coder assignments, reliability sample, adjudication log, analysis plan, frozen dataset, analysis code, or traceability matrix has been received.
No privacy, security, ethics, accessibility, measurement, legal, disclosure, or independent methods review is recorded.

State history

Effective time	State	Authority	Evidence
2026-08-31T00:00:00-04:00	`claim_received`	Founding steward intake	Request in the project workstream; no underlying study artifacts

No later state is implied. State advances are append-only and must satisfy the commissioned interview intake protocol.

Unregistered prospective candidate FDE-JA-002

The Prospective Job-Analysis Preregistration uses provisional ID FDE-JA-002 for design work. It is not a registered study entry and has no study-state authority.

Field	Current candidate state
Purpose	Prospective job analysis and content-blueprint evidence
Protocol	Draft registration candidate
Permanent registration / URI	None
Named research and operations roles	Unassigned
Ethics/privacy determination	None recorded
Sampling frame and numeric target	Unresolved
Recruitment or data-collection authority	None
Participant records	None
Findings or public claims	None

The unsigned FIELDWORK_AUTHORIZATION_TEMPLATE records its fail-closed operational state. A future machine register entry must be created only after the study ID and protocol are frozen; it must not backdate prospective registration or collection authority.

State transition authority

State entered	Required receipt
`claim_received`	Claim wording, source, date, intended public use, and explicit unverified state
`evidence_requested`	Named custodian, scoped artifact request, approved transfer path, and deadline
`evidence_received`	Package manifest, transfer receipt, malware/security check, access log, and preliminary gap list
`chain_of_custody_verified`	Custodian/collector attestations, digest verification, provenance review, and unresolved discrepancies
`record_validation_complete`	Schema-valid records, deduplication, consent/eligibility/quality decisions, and exclusion flow
`coding_complete`	Frozen codebook, coder training/assignments, double-code sample, agreement, and adjudications
`analysis_frozen`	Preregistered or explicitly retrospective analysis plan, dataset/code/environment digests, and frozen outputs
`independent_review_complete`	Signed independent methods, privacy, ethics, measurement, and disclosure reviews with disagreements
`aggregate_substantiated`	Machine-reproduced claim counts, no blocking record states, disclosure check, and signed aggregate receipt
`publication_authorized`	Exact public wording, methods package, limitations, citation, correction channel, and governance approval
`published`	Permanent release URI/archive, immutable manifest, first-publication time, and monitoring owner
`corrected`	New version, materiality decision, affected claims, remediation, and links to superseded release
`retracted`	Retraction authority, reason class, affected claims, participant remediation, and retained audit trail
`suspended`	Incident/defect, scope, hold authority, affected records/claims, and remediation decision
`withdrawn`	Authorized termination reason, data disposition, participant obligations, and public-status decision

Maintenance

The Research Director owns the register; the Claims Reviewer controls public wording; neither may approve their own evidence package alone.
Every claim and state has a stable identifier. Corrections append; they do not overwrite.
Aggregate numbers are generated from eligible record manifests. Hand-entered totals are display caches only and never authority.
Any page that cites a study must resolve its study ID, publication version, evidence cutoff, data-availability statement, and correction status.
A quarterly no-change check is not a study-state transition. Material new evidence opens a state event immediately.