Worked interpretation cases
Economic interpretation cases
Six source-based cases illustrate the definitions, publication context and sources required to assess economic measurements and research findings.
Standards and interpretation requirements
SDMX (Statistical Data and Metadata eXchange) can carry observations, definitions and source metadata. UCEM (Unified Code for Economic Measurement) specifies the economic meaning and sources an application should preserve when answering a question. Equivalent rules in an SDMX application can reach the same judgments.
These six curated teaching examples reuse retained publisher sources. The readings are authored explanations, not live model outputs or new semantic-engine results. The trade interpretation remains proposed. Source editions, locators and limits are inside each card.
01 · Output
GDP price basis
United States · 2024 Q4 · third estimate
- Nominal · current prices
- 29,723.9billion USD
- Real · chained 2017 dollars
- 23,542.3billion chained 2017 USD
Can either value answer a request for real GDP?
Reading with context
Use the chained-dollar estimate for real GDP. The current-dollar estimate has a different price basis.
Inspect source, meaning and limits
United States, 2024 Q4, third estimate. Both are seasonally adjusted annual-rate levels. A shared subject and quarter do not make their valuations interchangeable.
2025-03-27 · third estimate
page 9: Table 3, line 1, 2024 Q4, current_dollars; page 9: Table 3, line 1, 2024 Q4, chained_2017_dollars
Retained source fingerprint
4ad66187c3481e053b503ebdfe5c0c8466974e89dce4b8cf0ffb913e5117f73eThese are retained historical estimates, not current values. This reading performs no nominal-to-real conversion and does not establish complete historical methodology.
Requirements illustrated
Inspect the GDP record02 · Publication
Publication revisions
United States · 2024 Q4 · January and March 2025 releases
- Advance estimate
- 29,700.6billion USD
- Third estimate
- 29,723.9billion USD
Did GDP grow between the January and March publications?
Reading with context
These publications estimate the same 2024 Q4. Their different values describe a revision, not growth between two economic periods.
Inspect source, meaning and limits
Advance estimate published 30 January 2025; third estimate published 27 March 2025. Observation period and publication date have different roles.
2025-01-30 · advance estimate
page 9: Table 3, line 1, 2024 Q4, current_dollars
Retained source fingerprint
65cdc63c9f33ec1ec0933c369494bf99e7f549c3381c9095d52a51e69b7d7af12025-03-27 · third estimate
page 9: Table 3, line 1, 2024 Q4, current_dollars
Retained source fingerprint
4ad66187c3481e053b503ebdfe5c0c8466974e89dce4b8cf0ffb913e5117f73eOnly supplied editions are compared. This does not establish complete archives, an unchanged historical method or what was available at an earlier cutoff. No growth rate is calculated.
Requirements illustrated
Follow the publication editions03 · Prices
Index levels and inflation rates
U.S. city average · September 2025 · seasonally adjusted
- September 2025 CPI-U
- 324.245index · 1982-84=100
Does this mean prices rose by 324.245% this month?
Reading with context
No. This is an index level. A monthly inflation rate needs a specified comparison month and compatible index levels.
Inspect source, meaning and limits
CUSR0000SA0: all items, U.S. city average, all urban consumers, seasonally adjusted. The index reference and comparison interval are separate choices.
Retained 10 October 2026 bulk snapshot; original publication edition unresolved
line:946
Retained source fingerprint
47507ab13d9366d6e9f6a85a50bbd863179c97fa51c92ecb737875455d6ddc0fRetained 10 October 2026 series metadata
CUSR0000SA0 series row; base_period 1982-84=100
Retained source fingerprint
f7367ea7851338d4d9165061bfb14051030f07f073dc12d007cdff2becd2320dNo percentage change is calculated here. Original publication editions and times are unresolved; the retained bulk snapshot is not the original monthly release.
Requirements illustrated
Inspect the CPI record04 · Availability
Unavailable observations
U.S. city average · October 2025 · seasonally adjusted
- October 2025 CPI-U
- -publisher marker · footnote X
Can this observation be filled with zero?
Reading with context
Keep the observation unavailable. The dash is a publisher marker. It does not report a measured zero or establish zero inflation.
Inspect source, meaning and limits
CUSR0000SA0, seasonally adjusted, October 2025. The retained source keeps the marker, its footnote and the missing-value interpretation together.
Retained 10 October 2026 bulk snapshot; original publication edition unresolved
line:947
Retained source fingerprint
47507ab13d9366d6e9f6a85a50bbd863179c97fa51c92ecb737875455d6ddc0fThis example concerns the selected retained observation. It neither estimates the missing value nor claims that all publisher routes or later editions are unavailable.
Requirements illustrated
Inspect availability and sources05 · Trade
Trade valuation conventions
U.S. Census goods imports · Mexico row · December 2024
- Imports, customs basis
- 39,224.9million USD
- Imports, C.I.F. basis
- 39,541.1million USD
Can these two import values be substituted because both use USD?
Reading with context
Keep customs and C.I.F. values distinct. C.I.F. (cost, insurance and freight) includes import charges that customs valuation excludes.
Inspect source, meaning and limits
U.S. Census goods imports in the Mexico row, December 2024, not seasonally adjusted. Preserve the reporter, counterpart, period and valuation labels.
5 February 2025 · CB 25-17 / BEA 25-04
PDF p59, supplement p7, Exhibit 4, Mexico** row, Imports Customs basis, December; PDF p59, supplement p7, Exhibit 4, Mexico** row, Imports C.I.F. basis, December; valuation notes pp9-10
Retained source fingerprint
abbad725471e0e43b772fea178e9662086cc5a1ab456edc72596823008415153This is an informative proposed interpretation from the retained trade dossier, not an accepted trade profile. Exact historical method/classification identity remains unresolved. No trade balance or conversion is computed.
Requirements illustrated
Read the retained trade dossier06 · Research
Research sample counting units
Empowering Inclusive Work · Table A.6, column 1
- Column workers
- 43,605workers
- Column maximum
- 480,070worker-weeks
Does 480,070 establish the exact regression sample size?
Reading with context
No. It is the column maximum worker-weeks. The exact outcome-specific regression N remains unresolved.
Inspect source, meaning and limits
A working-paper appendix reports 43,605 workers and a maximum of 480,070 worker-weeks in Table A.6. The footer says N varies by outcome; counting units and scope matter.
Working-paper appendix · edition unresolved; retained 7 October 2026
Table A.6, column 1 footer and note, physical PDF page 14
Retained source fingerprint
20bc9f53a76e56dfa9eb2d6c62c38978a7f18cd26a692ff77071dc3150ea1010Exact appendix edition and matching main-paper version are unresolved. These are counts for one study, not population employment statistics or independent replications.
Requirements illustrated
Inspect the full research walkthroughSupplementary research analysisOpen the complete research walkthroughInspect the shared SDMX package, normalization example and five original interpretation questions.
Source publication
Empowering Inclusive Work
A retained NBER-published web appendix studying AI and deaf or hard-of-hearing workers. Selected sources: Tables A.6 and A.9, physical PDF pages 14 and 28.
Original capture: 7 October 2026. Exact appendix edition and matching main-paper version remain unresolved. This view covers selected cells, not the complete paper.
Open the NBER paper page- Baseline attrition coefficient
-0.210- Its reported standard error
0.052- Exact regression sample size
- Unresolved
Coefficient construction and interpretation
Does −0.210 mean a 21-percentage-point reduction?
A number alone cannot distinguish an unscaled probability difference from a normalized coefficient. The definition and its source are needed to answer this question.
Deliberately incomplete teaching view
Value without construction metadata
-0.210
- Coefficient construction
- Omitted from this view.
- Scale and normalization
- Omitted from this view.
The number alone is insufficient.
Calling it a 21-percentage-point reduction would introduce an unsupported assumption about the coefficient's scale. Retrieve the definition before interpreting it.
Direct A.9 sources retained
Value with construction metadata
-0.210
- Coefficient construction
- OLS coefficient divided by the pre-AI DHH mean: the mean for deaf or hard-of-hearing workers before AI introduction.
- Source locator
- A.9, Panel C, column 2 and note; physical PDF page 28.
A 21-percentage-point interpretation is unsupported.
The source labels this coefficient as normalized. Preserve that basis when reporting it. No conversion or causal validation is performed here.
Missing information and missing rules are different problems. This omission is authored for illustration; it is not the SDMX baseline below. The complete shared SDMX packet already carries this context, and a reader can answer from it.
UCEM's documentation obligations retain construction, scale, sources and unresolved states, and require disclosure of unsupported interpretations. The existing RE03, RE04 and RE06 obligations describe those requirements. An application can implement equivalent rules without using UCEM.
This explanation reads A.9 directly. Matching values do not establish A.9's binding to the A.6 baseline cell; that correspondence remains a candidate interpretation.
Inspect the worked scale question and its sourcesOne source package, available to both views
Context in the SDMX representation
This is OpenUCEM's existing local experimental SDMX 3.1 / SDMX-JSON 2.1.0 representation, not a native NBER SDMX feed. It includes data, structures, reference metadata, layout bindings and a typed context attachment. No information is removed from the SDMX view.
| Quantity role | Retained packet token | Meaning boundary |
|---|---|---|
COEFFICIENT | -0.210 | Reported coefficient; keep its normalization context. |
STANDARD_ERROR | 0.052 | A.6 explicitly reports worker-clustered standard errors. |
WORKERS | 43605 | Column workers, not an established outcome-specific count. |
MAX_WORKER_WEEKS | 480070 | Column maximum worker-weeks, not exact regression N. |
NORMALIZATION_DENOMINATOR | 0.056 | Rounded printed denominator; no arithmetic performed. |
Coefficient and SE preserve the printed decimal tokens. Counts are integer transcriptions in the retained packet; the PDF footer prints 43,605 workers and 480,070 maximum worker-weeks.
Tables may scroll sideways. Focus the table region and use the Left and Right arrow keys.
Task-specific source documentation requirements
Assessment of proposed claims
Open a question to compare the scopes. Package validation checks the data and its references. The added UCEM checks apply disclosed, case-specific source documentation obligations to a proposed answer. Labels below are editorial, not new UCEM dispositions or an API.
The build validates and decodes the packet, checks the fixed exact-N and inapplicability premises, and compares the proposed SE with the selected cell. Explanatory rules and conclusions are authored; they are not inferred from the question text. Question 2 reads A.9 directly and does not establish its binding to A.6.
01Report the finding with its limitsWhat does the selected baseline attrition cell report?
SDMX package checks
Complete structured sources
Package checks passed
Three official message schemas, typed context, local DSD/reference joins and reconstruction are checked at build time.
- Judgment on this answer claim
- Not evaluated by package validation.
The relevant values, qualifiers, gaps and source references remain available in the full packet. An application needs explicit domain rules to assess this question.
Same SDMX sources + UCEM task contract
Faithful partial report
The baseline cell reports -0.210, with standard error 0.052.
A.6 explicitly identifies worker-clustered standard errors. Its normalization binding to A.9 remains a candidate interpretation.
- Sources used
- A.6, Panel B, Attrition, column 1; PDF page 14
- A.6 table notes
- Existing obligations illustrated
- RE04RE05RE06DE03
Limits retained in this answer
- The source mapping remains proposed; this is a source-disclosure judgment, not causal certification.
- Exact appendix edition: unresolved.
- Calendar window: unresolved in selected A.6 source material.
- Complete controls and fixed effects: unresolved.
02Keep the normalization attachedDoes Table A.9 report a 21-percentage-point reduction in attrition?
SDMX package checks
Complete structured sources
Package checks passed
Three official message schemas, typed context, local DSD/reference joins and reconstruction are checked at build time.
- Judgment on this answer claim
- Not evaluated by package validation.
The relevant values, qualifiers, gaps and source references remain available in the full packet. An application needs explicit domain rules to assess this question.
Same SDMX sources + UCEM task contract
Unsupported percentage-point interpretation
The supplied sources do not establish a 21-percentage-point change.
A.9 labels -0.210 as an OLS coefficient divided by the pre-AI DHH mean. Keep that normalization; this checker performs no conversion.
- Sources used
- A.9, Panel C, column 2 and note; PDF page 28
- Existing obligations illustrated
- RE03RE04RE06
Limits retained in this answer
- The source mapping remains proposed; this is a source-disclosure judgment, not causal certification.
- Exact appendix edition: unresolved.
- Calendar window: unresolved in selected A.6 source material.
- Complete controls and fixed effects: unresolved.
03Do not turn a maximum into exact NDoes Table A.6 establish exactly 480,070 observations for baseline attrition?
SDMX package checks
Complete structured sources
Package checks passed
Three official message schemas, typed context, local DSD/reference joins and reconstruction are checked at build time.
- Judgment on this answer claim
- Not evaluated by package validation.
The relevant values, qualifiers, gaps and source references remain available in the full packet. An application needs explicit domain rules to assess this question.
Same SDMX sources + UCEM task contract
Exact regression N unestablished
The column maximum is 480,070 worker-weeks. It is not an established attrition-regression N.
N varies by outcome; A.6 points to main Table 2. Footer is a maximum, not an exact per-outcome N.
- Sources used
- A.6 column 1 footer and table notes; PDF page 14
- Existing obligations illustrated
- RE05RE06
Limits retained in this answer
- The source mapping remains proposed; this is a source-disclosure judgment, not causal certification.
- Exact appendix edition: unresolved.
- Calendar window: unresolved in selected A.6 source material.
- Complete controls and fixed effects: unresolved.
04Inapplicable does not become zeroDoes Table A.6 column 5 estimate zero attrition effect?
SDMX package checks
Complete structured sources
Package checks passed
Three official message schemas, typed context, local DSD/reference joins and reconstruction are checked at build time.
- Judgment on this answer claim
- Not evaluated by package validation.
The relevant values, qualifiers, gaps and source references remain available in the full packet. An application needs explicit domain rules to assess this question.
Same SDMX sources + UCEM task contract
Inapplicable, with an explicit reason
Column 5 supplies no attrition estimate; it does not report zero.
Selection requires presence before and after AI; source does not analyze attrition in this selected group.
- Sources used
- A.6, Panel B, Attrition, column 5 and note; PDF page 14
- Existing obligations illustrated
- RE04RE05RE06
Limits retained in this answer
- The source mapping remains proposed; this is a source-disclosure judgment, not causal certification.
- Exact appendix edition: unresolved.
- Calendar window: unresolved in selected A.6 source material.
- Complete controls and fixed effects: unresolved.
05Keep each estimate with its own uncertaintyCan the baseline attrition coefficient use the delivery-time standard error 0.004?
SDMX package checks
Complete structured sources
Package checks passed
Three official message schemas, typed context, local DSD/reference joins and reconstruction are checked at build time.
- Judgment on this answer claim
- Not evaluated by package validation.
The relevant values, qualifiers, gaps and source references remain available in the full packet. An application needs explicit domain rules to assess this question.
Same SDMX sources + UCEM task contract
Wrong estimate and uncertainty pairing
The selected attrition coefficient -0.210 has reported SE 0.052.
The proposed SE 0.004 does not match that cell. Delivery time is a separate outcome and regression.
- Sources used
- A.6, Delivery Time and Attrition rows, column 1; PDF page 14
- Existing obligations illustrated
- RE03RE04RE06
Limits retained in this answer
- The source mapping remains proposed; this is a source-disclosure judgment, not causal certification.
- Exact appendix edition: unresolved.
- Calendar window: unresolved in selected A.6 source material.
- Complete controls and fixed effects: unresolved.
Scope of the comparison
An SDMX application with equivalent domain rules can reach the same judgments. This example applies information that is already in the SDMX packet. It demonstrates five source-bound checks, not a general semantic engine, autonomous agent or format advantage.
UCEM specifies bounded meaning and source documentation obligations for humans and agents discovering, interpreting and using economic sources. Publisher taxonomies and existing standards remain foundations. EQUA remains an earlier research proposal, with no reserved engine or publication-protocol implementation stage.
Download the worked requests, results and source-review limitsSource and review scope
A fresh reviewer read only the retained PDF pages, title and metadata, with the author's key and questions withheld. They confirmed the selected tokens and count/inapplicability distinctions. This is bounded source verification on previously used material, not measured extraction accuracy or causal validation.
A.9 does not itself identify its parentheses as standard errors or specify clustering. This demonstration uses A.6 for SE kind/clustering and A.9 for its explicit normalization label. Correspondence across the tables remains scoped interpretation.
The full publisher PDF is not redistributed in the download. Current web retrieval was unsuccessful; this view uses the retained capture. Hashes identify bytes, not publisher authenticity or economic truth.
- Retained appendix SHA-256
20bc9f53a76e56dfa9eb2d6c62c38978a7f18cd26a692ff77071dc3150ea1010- Source mapping authority
- Proposed interpretation. This demonstration adopts no new normative profile or operation.
- Website search coverage
- This separate worked comparison is outside the three-item example search inventory.
Applicable specification requirements
Each example links its source and the bounded obligations it illustrates. Source checks, economic interpretation and permission to calculate have separate scopes.
Read the specification