Collection scope and limitations
Collection coverage
The three curated records, the discovery pilot and the interpretation cases have separate inventories. They reuse some sources; their counts must not be added as independent findings or presented as complete publisher coverage.
| Record | Exact included scope | Material limitations |
|---|---|---|
| BEA GDP | Six 2024 Q4 cells: nominal and real across advance, second and third estimates. | Historical publications; complete method identity and archive coverage unestablished. |
| BLS CPI | 24 monthly records across CUSR0000SA0/CUUR0000SA0, September 2025 through August 2026. Two October observations unavailable. | Captured bulk history; original publication editions/times unestablished. M13 annual averages excluded. |
| NBER 35814 | One employment-to-population expression: Table 3, 2021 column, physical PDF page 70. | Working paper; summary-cell regression N and exact conditioning unresolved. Other estimates excluded. |
24 records in the discovery pilot
Discovery search covers 20 selected research cells from two NBER studies and four historical BLS 2024 annual-average statistics. The research inventory contains 19 numeric estimates and one inapplicable cell. Supporting context is not another independent finding. Ranking uses keywords and explicit metadata filters.
The BLS archive was inspected through a publisher HTML view after a native download returned HTTP 403. The retained manual transcription is not an original publisher-byte capture. Those broad disability statistics do not establish hearing-specific employment or an AI treatment effect. NBER edition, sample and scale uncertainties remain attached.
Interpretation cases and supplementary research analysis
The six-example comparison reuses selected GDP, CPI, trade and research sources. Its cards are authored explanations, outside the three-item search and the 24-record discovery collection. The trade mapping remains informative and proposed; no new arithmetic or model performance is demonstrated.
The optional research walkthrough retains five original questions using selected AI-and-disability sources from Tables A.6 and A.9. It supplies no new independent findings or measured agent-performance result. Both checking views retain the same complete SDMX sources. A.9's normalization does not establish its exact binding to A.6.
Interpretation of search results
A no-match on the retained-reference search concerns its three curated records. The public library also searches saved submissions within its stated coverage. A no-match in the discovery pilot concerns its 24 retained records and the applied keywords/filters. Neither establishes absence from an agency, journal or the wider literature. Truncated candidates and missing source information require separate disclosure.
Publication time, capture time and interpretation review time remain separate. A current source link may show a later edition. Retained hashes identify inspected bytes, with the explicitly disclosed exception of the BLS HTML transcription.
Source policy and application scope
These collections use direct publisher sources and reviewed editorial projections. Data Commons is solely a conceptual and technical reference; none of its economic observations or mappings supplies these records. The worked collections do not constitute live publisher acquisition. Publication intake is a separate, user-requested publication-analysis function whose availability depends on provider configuration. Its source coverage and proposed interpretations remain explicit.