Canonical empirical sources catalog and platform-wide derivation convention

v1.0 · Created May 7, 2026 for v3.2.8 (mitigation of Comprehensive Verification Report Dimension 5 finding: documents could benefit from explicit source citations on derivational claims) · Jason Robertson · Ohio · 2026 · content last changed v3.8.457

SWEPT AND CURRENT (v3.8.362). This document was reviewed line by line at v3.8.360, every numerical claim in it checked against the registry and the models, and the corrections applied in place. Its figures are current as of v3.8.361. THE NOTICE THAT FOLLOWS IS KEPT AS HISTORY, not as a live warning: it was written before that review and is superseded by it. Where a claim could not be restated -- because the finding itself was overturned rather than the number -- the document says so at the point of use. Original notice: UNDER REVISION (v3.8.337). This document's figures for the Social Security transition are superseded. Measured in today's dollars, with the workers the phase-out grandfathers actually retiring and claiming, phasing out Social Security alone would require about $58 trillion of cumulative borrowing by 2085, and the platform's reform reduces that to about $35 trillion: a reduction of about 40 percent, not the 99.9 percent (from $93.9 trillion to $82 billion) or the peak of about $1.9 trillion stated in earlier versions. Passages that rest on those figures will be rewritten; see the version log for v3.8.336 and v3.8.337. UPDATED v3.8.341: both models have since been rebuilt on the decided repayment rule -- workers leave oldest first and get back their pre-enactment contributions with interest at the government's own borrowing cost, before they retire. On that basis the standalone sunset needs about $73 trillion of cumulative borrowing by 2085 and the platform's reform about $60 trillion, a reduction of about 18 percent rather than the 40 percent stated when this notice was written, or the 99.9 percent stated in the document below. UPDATED v3.8.347: the platform has since adopted ABSORPTION, which it always intended -- the Social Security trust fund moves into the Sovereign Fund at year 1 and every payroll tax dollar of the wind-down goes to the fund, with benefits and repayments financed by borrowing. On that design the fund reaches about $137 trillion by year 60 and the platform REDUCES the federal deficit by about $2.72 trillion a year at its published 4 percent return, or $1.39 trillion counting interest on the transition debt, which stands at about $85 trillion. The gain is a spread between the fund's return and the government's borrowing cost and reverses below about 3 percent. Separately, the pillars' own build-out years are now costed for the first time: about $0.321 trillion [UPDATED v3.8.414: the first-year figure moved to $0.321 trillion when the pillar ledger adopted the CMS 2024 column at v3.8.409 -- more existing spending absorbed means a smaller first-year deficit] in year 1, peaking near $1.7 trillion in year 6, roughly $20 trillion over fifteen years, none of it inside the Social Security figure. See 05_Transition_Design_Options_And_Costs.docx and 05_Federal_Fiscal_Impact_Analysis.docx, which carry the current figures.

Purpose

This document catalogs the canonical empirical sources that platform analytical documents draw from, and establishes the platform's sourcing convention so readers can trace numerical claims back to their underlying sources. The Comprehensive Verification Report at v3.2.7 found that approximately forty percent of numerical claims in analytical documents do not have source-indicator language within two hundred fifty characters, but inspection showed most are derivational (e.g., a claim like '$67 billion equals the exemption above the standard deduction summed across 131.9 million returns' is calculation from sourced inputs rather than a separately citable empirical claim). This document addresses that finding by making the platform's sourcing taxonomy explicit and providing a single canonical reference that all analytical documents can point at.

Sourcing Convention

[FOURTH LIMB ADDED v3.8.360. The convention below describes three paths, and the platform's practice has outgrown them. Many figures are now (d) RECOMPUTED BY A NAMED TOOL FROM SEALED SOURCE DATA, AND CHECKED ON EVERY RELEASE. tools/build_pillar_ledger.py computes the platform's totals pillar by pillar rather than restating them; tools/verify_internal_consistency.py RECOMPUTES registered figures and fails when one drifts; tools/fund_ledger.py, tools/transition_lever_model.py and tools/montecarlo_bridge_sustainability.py each carry a self-check against the workbooks they mirror; and the audit's ANALYSIS-TOOL-DRIFT, ANNOTATION-STALE, PILLAR-LEDGER-DRIFT and NUMERIC-COVERAGE checks fail a release when the tools, the annotations, the ledger or the documents fall out of step. THIS IS A STRONGER GUARANTEE THAN CITATION: a cited figure is as good as the reader's willingness to follow the citation, while a recomputed one is checked whether anyone looks or not. A reader tracing a claim should expect one of four paths, not three.] [TERMS THAT MEAN TWO THINGS, ADDED v3.8.397. 'ABSORPTION' is used in this platform for two unrelated things, and a reader who meets the word without knowing which one is meant will misread the arithmetic. FUND ABSORPTION is the design choice at ASSUM-transition-funding-design, modelled at v3.8.343: the OASI trust fund moves into the Sovereign Fund at year one, the payroll tax of the wind-down compounds there, and benefits are financed by borrowing. It is why the platform reduces the federal deficit at maturity and why the wind-down borrows about $85 trillion by design. PROGRAMME ABSORPTION is the accounting total at FIG-existing-programs-absorbed: about $2.12 trillion a year of existing federal spending the pillars replace, which is what turns $4.96 trillion of commitments into $2.83 trillion of net new spending. THE TWO NEVER INTERACT. Where a document says only 'absorption', check which is meant before using the number.] The platform's analytical documents follow this convention: numerical claims are either (a) directly cited to a primary source at the point of use, (b) derived through explicit calculation from inputs cited elsewhere in the same document, or (c) derived from the canonical sources cataloged below and referenced at the top of each analytical document. Derivational claims of the form '$X equals Y times Z' or 'X percent of population Y equals N people' inherit their sourcing from the inputs to the calculation; the calculation itself is the source. Readers can verify any platform claim by tracing it back through one of these three paths.

Each analytical document includes a Sources baseline statement near its top identifying which entries from this catalog are relevant to that document's claims. The catalog below enumerates the canonical sources and identifies which documents draw from each.

Fixed Analytical Baseline

The platform's quantitative claims are anchored to a fixed analytical baseline — a specific vintage of each external source, held constant so that figures across documents stay mutually consistent and reproducible. Two sources warrant explicit note. For healthcare, the model is anchored to CMS National Health Expenditure data for 2023 (the $15,475 [UPDATED v3.8.406: the registered healthcare baseline moved from the CMS 2023 release to CMS 2024, $14,570 = $15,475 per capita, which was already sealed in the package's own data archive. A verifier case now checks it against that source on every build] [CORRECTED v3.8.294: read $15,475, ten dollars below the sealed 2023 CMS value] per-capita baseline and the reduction-target derivation); the 2024 release shows per-capita spending has since risen to $15,474 (18.0 percent of GDP), which the platform reports as the current cost trend while retaining the 2023 figure as the conservative model baseline, so the reduction and savings estimates are if anything understated. For the wage floor, the anchors use Bureau of Labor Statistics Occupational Employment and Wage Statistics data, rolled forward to May 2025 (the most recent release); the wage figures remain illustrative magnitudes, as a production implementation would pull the full OEWS catalog at the metropolitan-area level. The platform adopts newer releases only through deliberate, versioned roll-forward passes that recompute or re-anchor every dependent figure together, rather than piecemeal, so the baseline never drifts mid-analysis. The complete vintage list for every source appears in the Canonical Sources Catalog below.

Canonical Sources Catalog

Federal Tax and Income Sources

Internal Revenue Service (IRS) Statistics of Income (SOI) 2023 baseline: source for individual filer counts, adjusted gross income distributions, and tax-bracket population data. Used in: Federal Income Tax Revenue Modified Architecture; Combined Reform Model income tax sheets; Federal Fiscal Impact Analysis tax-revenue stream; What This Means For You household examples; Per Citizen Benefits and Costs.

IRS Form 1040 instructions and Publication 17 (2024 tax year): source for current standard deductions ($14,600 single / $29,200 MFJ / $21,900 HOH), tax brackets, child tax credit ($2,000 per qualifying child), and other current-system parameters used as the comparison baseline. Used in: We The People Calculator; Federal Income Tax Revenue Modified Architecture; Narrative Example 100K Tax Comparison. The comparison baseline is deliberately anchored to the tax-year-2024 federal code for analytical stability across the platform's tax examples. The One Big Beautiful Bill Act (OBBBA, enacted July 2025) has since changed several of these current-system parameters, effective tax year 2025: the Child Tax Credit rose to $2,200 per qualifying child; standard deductions rose to $15,750 single / $31,500 married-filing-jointly (2025) and $16,100 / $32,200 (2026); a temporary senior deduction and deductions for tip and overtime income were added; and the Social Security wage base rose to $176,100 (2025). Because OBBBA made the current system more generous, the platform's comparisons against the 2024 baseline modestly overstate the platform's relative advantage; a full roll-forward of the tax examples to the post-OBBBA code is tracked as RESEARCH-22 in the Tracked Issues Registry.

U.S. Treasury and Joint Committee on Taxation revenue projections: source for current federal individual income tax aggregate revenue (approximately $2.4 trillion annually). Used in: Federal Income Tax Revenue Modified Architecture; Wage Floors As Tax Architecture; Federal Fiscal Impact Analysis.

Healthcare Sources

Centers for Medicare & Medicaid Services (Centers for Medicare and Medicaid Services (CMS)) National Health Expenditure (NHE) data 2023: source for U.S. healthcare spending per capita ($15,475 baseline used throughout the platform), aggregate healthcare spending, and spending breakdowns by payer. Used in: Healthcare Transition Detailed Plan; Pillar Four Universal Healthcare Access; Federal Fiscal Impact Analysis healthcare stream; Per Citizen Benefits and Costs; Path To Reality. The 2024 NHE release ($15,474 per capita, 18.0 percent of GDP) is reported as the current cost trend, with the 2023 figure retained as the model baseline.

Kaiser Family Foundation (KFF) employer health benefits surveys: source for average employer-sponsored insurance premiums and employee contributions. Used in: We The People Calculator current-system premium baseline; What This Means For You; Healthcare Transition Detailed Plan.

Demographic and Population Sources

U.S. Census Bureau population estimates 2023: source for total U.S. population about 337 million [UPDATED v3.8.360] fixed at 336,755,052 by the author's Decision B at v3.8.333, from the sealed Census source; this read '~334 million', household counts, and demographic breakdowns by age and household composition. Used in: per-capita calculations throughout; Federal Fiscal Impact Analysis denominators; Per Citizen Benefits and Costs. Note: the US resident-population reference figure is cited inconsistently across documents (approximately 320 million to about 337 million); standardizing it to the current approximately about 337 million (2024 Census) and re-deriving the population-dependent figures is tracked as RESEARCH-22.

Bureau of Labor Statistics (BLS) Current Population Survey and wage data 2023-2024: source for employed population (~158 million), wage distributions by occupation, hours worked, and labor force participation rates. Used in: Empirical Wage Floors substantiation; Gender Pay Gap analysis; Combined Reform Model labor inputs; Disemployment sensitivity analysis.

Sovereign Fund and Investment Sources

Norway Government Pension Fund Global (GPF-G) annual reports: comparative source for sovereign wealth fund management practices, real returns achieved over multi-decade horizons, and governance structures. Used in: Sovereign Fund Governance Design; How This Was Built; Combined Reform Model sovereign fund stream.

the sealed 2026 Social Security Trustees Report (_data_archive/2026-08-27_ssa_trustees_2026/): [UPDATED v3.8.360] this read 'Trustees Reports 2023-2025'. The platform moved to the sealed 2026 report at v3.8.333, and every payroll base, mortality table and economic assumption now comes from it. The 4.42 percent of taxable payroll and approximately $25 trillion present value stated below are from the earlier vintages and are NOT restated, because nothing in the platform currently uses them. Formerly: source for Social Security trust fund actuarial-deficit measures (4.42 percent of taxable payroll; approximately $25 trillion in present value), demographic projections, and trust fund depletion timelines. The $93.9 trillion [RECOMPUTED v3.8.295: read $63 trillion. That figure applied the trust fund's EARNING rate to the phase in which the government BORROWS. At the decided 4.0 percent borrowing rate -- anchored to CBO's sealed projection of the average rate on federal debt, mean 4.01 percent over 2026-2056 -- the same model gives $93.9 trillion. The reform side was recomputed too: its peak is $2.01 trillion and the reduction is 97.9 percent] figure cited throughout is a separate platform-derived cumulative (undiscounted) 60-year transition-financing estimate, not the SSA actuarial deficit. Used in: Community Contribution Plan White Paper; The Founding Stake; Two Paths Compared; Federal Fiscal Impact Analysis. The 2026 Trustees Report (released June 2026) moved OASI trust-fund depletion one year earlier, to 2032 (fourth quarter), with 78 percent of benefits payable thereafter (combined OASDI depletion moved to 2034); the 75-year actuarial deficit rose to 4.42 percent of taxable payroll (from 3.82 percent), driven largely by the 2025 reconciliation act reducing income-tax revenue on benefits and a lower assumed fertility rate, so the $93.9 trillion comparison baseline is anchored to the 2024 report and remains conservative.

Infrastructure and Civic Sources

American Society of Civil Engineers (ASCE) Infrastructure Report Card 2025: source for infrastructure investment gap estimates (the $1.2 trillion highways and bridges figure, $140 billion transit gap, and other infrastructure shortfalls). Used in: Physical Civic Infrastructure Substantiation; Federal Infrastructure Fee; Civic Infrastructure Architectural Framing.

Federal Communications Commission (FCC) and Bureau of Industry and Security broadband data: source for current residential broadband market revenue (approximately $111 billion annual, 120 million subscribers, average $77 per month). Used in: Free Universal Broadband Cost Analysis; Universal Broadband Access Substantiation.

Federal Trade Commission (FTC) Consumer Sentinel Network identity-theft reports: source for identity-theft cost estimates (the $8 billion credit account fraud figure, $3 billion government benefits fraud figure, and aggregate consumer harm). Used in: Identity Theft Reduction; Modernize Civic Engagement Integrated Argument.

Family and Care Sources

FAMILY Act modeling and state paid family leave program data (California PFL, New Jersey FLI, New York PFL, Washington PFL): source for paid family leave cost projections, utilization rates, and contribution rate calibration. Used in: Pillar Eight Universal Paid Family Time; Federal Fiscal Impact Analysis Pillar Eight stream.

Quebec $7-day childcare program evaluations and U.S. childcare cost surveys: source for universal childcare cost projections at U.S. scale and the $180 billion aggregate cost figure. Used in: Universal Childcare Pillar; Federal Fiscal Impact Analysis childcare stream.

Long-term care expenditure data from CMS and American Association of Retired Persons (AARP)/Genworth Cost of Care surveys: source for the $475 billion aggregate long-term care expenditure figure [SOURCED v3.8.273: the CMS component of this aggregate is now archived and reads $389.3 billion for 2024 -- home health $169.4B (NHE Table 14) plus nursing care facilities $219.9B (Table 15), both sealed at 2026-08-27_cms_nhe_2024. THE REMAINING ~$86 BILLION IS THE PART ATTRIBUTED TO AARP AND GENWORTH AND IS NOT SOURCED; the source register warns that the AARP figure is an imputed value that must not be summed with expenditures]. Used in: Aging In Place Implications.

Internal Platform Documents As Source

U.S. Government Accountability Office (GAO) federal IT reports (GAO-25-107795; GAO-23-106821): source for federal information technology spending (more than $100 billion per year, of which approximately 80 percent is operations and maintenance of existing and legacy systems).

Congressional Budget Office (CBO) and Congressional Research Service (CRS R48472) Highway Trust Fund analyses: source for federal surface-transportation funding (Highway Trust Fund highway-account outlays of approximately $50 billion in FY2023, primarily fuel-tax revenue supplemented by general-fund transfers).

U.S. Digital Service 2024 Impact Report: source for USDS impact metrics (18.25 million Veterans served and VA trust at 79.3 percent; a 130 percent increase in Affordable Connectivity Program enrollment to 23 million households; an estimated $285 million in projected five-year Social Security Administration infrastructure savings; and 749 million COVID-19 tests delivered).

Congressional Research Service (CRS IN12493), Executive Order 14158 (January 20, 2025), and Office of Personnel Management statements (November 2025): source for the U.S. Digital Service history and its 2025 reorganization as the U.S. DOGE Service and subsequent loss of centralized status.

The Combined Reform Model (04_Combined_Reform_Model.xlsx) is itself the canonical source for many platform claims, including the $122 trillion [CORRECTED v3.8.234. This figure is GROSS OF EVERY PILLAR DISBURSEMENT: the workbook it comes from nets out none after year nineteen. The platform now publishes a band computed by tools/fund_ledger.py across its own 4-7 percent return range and leads with the low end. [UPDATED v3.8.360] that band read '$60 trillion at the low end, $73 trillion in the middle, $113 trillion at the top' and is superseded twice over, by the today's-dollars rebuild and then by absorption: it now runs from about $137 trillion at a 4 percent return to about $345 trillion at 6 and $559 trillion at 7. Reproduce: python3 tools/fund_ledger.py --return 0.04. The gross figures are kept as history rather than deleted.] 60-year sovereign fund corpus projection, the about $137 trillion at a 4 percent return, in today's dollars 4-percent-return parallel scenario, and the year-by-year fiscal trajectories. The model is sourced from the inputs cataloged above; downstream documents that cite the model are inheriting that sourcing chain.

The Federal Fiscal Impact Analysis (05_Federal_Fiscal_Impact_Analysis.docx) consolidates the platform's fiscal claims and serves as the canonical source for aggregate-level pillar costs ($180 billion childcare, $145 to $200 billion (multi-method central ~$145B; $200B conservative upper anchor) mental health, $40-60 billion paid family time, etc.). Downstream documents that reference these aggregate figures are inheriting the FFIA's sourcing chain.

The Open Issues Registry (09_Open_Issues_Registry.docx) explicitly catalogs every empirical claim that requires external credentialed validation as RESEARCH-1 through RESEARCH-45 plus PERSONA-SIG-3, 4, 5. [UPDATED v3.8.360] this read 'RESEARCH-1 through RESEARCH-8'; four were added during the v3.8.348-358 document sweep, including the stress-sheet rebuild (RESEARCH-44) and the wage-floor revenue cost (RESEARCH-45). Claims tracked in those Section 47 entries are explicitly NOT certified by the platform; they are documented as positions requiring validation through the External Engagement Plan and Academic Outreach Letter Templates.

Assumptions Register

Every claim in this platform rests on one of two kinds of input, and this section makes the boundary between them explicit. Empirically-anchored inputs are drawn from referenced real-world sources — for example, the wage floor from the Bureau of Labor Statistics Occupational Employment and Wage Statistics program and national health expenditure from the Centers for Medicare and Medicaid Services — and are reproducible from those sources. Modeling assumptions are the forward-looking values the analysis must adopt where no single observed figure settles the question: rates of return, behavioral responses, long-run cost and benefit projections, and the trip points of the rollout’s circuit breakers. A sixty-year fiscal projection cannot be free of the second kind; the discipline this platform holds itself to is not the absence of assumptions but their full disclosure, bounding, and external validation.

To keep that disclosure honest and in one place, the platform maintains a machine-readable assumptions registry as the single source of truth for each modeling assumption. Every entry records the assumption’s current value or scenario, its plausible range or distribution where one is known, the source or basis for that range, the tracked-issue identifier under which it awaits validation, and the downstream claims that depend on it. An audit check verifies the registry’s internal integrity and that each entry points to a real open item.

The registry’s present state is itself a transparency statement. Most of the platform’s quantitative assumptions are recorded as awaiting credentialed external calibration rather than as settled figures, and the single most consequential of them is the Sovereign Fund’s long-run real rate of return, on which the sixty-year fund path and every bridge-financed pillar’s feasibility depend. The platform does not assert calibrated values it has not earned; the registry shows exactly where that line falls.

Each assumption also records the role it plays in the platform’s uncertainty analysis: a driver that is sampled directly from a calibrated distribution, a sensitivity parameter swept across a literature-informed range, an output whose distribution is propagated from upstream inputs rather than assumed, or a policy threshold that is calibrated against the driver distributions rather than guessed. The forthcoming stochastic analysis reads its distributions from the registry so that its assumptions are auditable rather than improvised. For the full surface of open design questions that are not parametric assumptions — unspecified mechanisms, depth-of-analysis gaps, and similar items — see the Open Issues Registry, which this registry cross-references rather than duplicates.

Acknowledged Limits

This convention establishes a sourcing taxonomy for derivational claims; it does not by itself validate the empirical defensibility of those claims. Full empirical defensibility validation requires credentialed external review. The platform's Open Issues Registry tracks the specific items requiring such review under RESEARCH-1 through RESEARCH-45 plus PERSONA-SIG items. [UPDATED v3.8.360 from 'RESEARCH-1 through RESEARCH-8'.] The External Engagement Plan documents engagement targets, reviewer profiles, and specific questions for each item. Until that external review is conducted and integrated, claims in the platform should be read with the understanding that they reflect rigorous internal analysis from cataloged sources but have not been validated by credentialed third-party review.

Within those limits, this convention enables readers to trace any platform numerical claim through one of three paths: direct citation at the point of use, explicit derivation from inputs cited in the same document, or inheritance from the canonical sources cataloged here. A claim that does not trace through one of these paths is itself a finding requiring documentation; readers encountering such claims should report them so they can be documented in the Open Issues Registry.

Cross-References

This document is the operational output of the Comprehensive Verification Report's Dimension 5 finding (recorded in Section 95 of the Open Issues Registry). It is referenced from the Sources baseline statements added in v3.2.8 to the platform's analytical documents. It complements the External Engagement Plan, which documents the path to credentialed external validation of empirical claims, and the Open Issues Registry's RESEARCH and PERSONA-SIG entries, which specify which claims explicitly require such validation.

Return-assumption convention (v3.7.661): the illustrative individual-outcome figures (approximately $1.23 million account, roughly $49,000 income, 1.40 times the current benefit) [UPDATED v3.8.360] THESE ARE SUPERSEDED: at the platform's published 4.0 percent lead the same inputs give about $865,000 and $34,600, roughly parity with the current average Social Security benefit rather than 1.40 times it. The corrected treatment is in 03_Community_Contribution_Plan_WhitePaper.docx and the $122 trillion [superseded v3.8.234: the published band is roughly $60-113T, led by $60T] 60-year fund corpus use an optimistic 5.5 to 6 percent real return. The sourced baseline is 4.28 percent (Norway GPFG realized, net of costs, since 1998), at which the corpus is approximately about $137 trillion at a 4 percent return, in today's dollars and the individual account is about $865,000 [RESTATED v3.8.370] from 'the individual figures are materially lower'. The individual-account return and the re-derivation of these figures at the baseline are tracked as RESEARCH-21 (pending external actuarial calibration); the fund return is RESEARCH-5.

Range-of-values convention. Where a figure depends on an assumption or input that cannot be pinned to a single independently-verified value (for example a scenario-dependent return, a pending-validation cost band, a target still under review, or a reference base with genuine estimate uncertainty), the platform presents that figure as a labeled range or band rather than a false-precision point estimate. The band endpoints and the driving assumption are stated explicitly, and the figure is linked to its tracked issue where one exists. This convention already governs entries such as the long-term-care commitment (a $200-400 billion net-new range), the Sovereign Fund corpus (a $62.5-122 trillion [superseded v3.8.234: the published band is roughly $60-113T, led by $60T] scenario range by return), and the individual retirement-account outcome (optimistic at 6 percent, closer to parity at the sourced 4.28 percent baseline, pending RESEARCH-21). A range is used specifically when a single point would assert precision the platform cannot yet defend; it is not a substitute for resolving the underlying question, which remains tracked for independent review.

Narrative-example lookup. Building on the range-of-values convention, each figure in the registry may carry a single labeled example value for use in prose, drawn from within its range: the conservative end for headline and aggregate figures (so claims read as floors) and the central value for individual pillars. Every example records its chosen end, rationale, citation, and validation status; figures still under review carry 'illustrative pending validation' with their tracked-issue reference. Prose uses the example value, while the full range, source, and status remain in the registry for the reader. An audit check enforces this structure and requires any pending example to cite its tracked item, so a precise-looking figure can never outrun its validation.

Robertson, J. (2026). Sources and Derivation Convention. We The People Platform (Version 3.8.484). https://wethepeopleplatform.com/_web_html/05_Analytical_Framing/05_Sources_And_Derivation_Convention.html

Robertson, Jason. 2026. "Sources and Derivation Convention." We The People Platform v3.8.484. https://wethepeopleplatform.com/_web_html/05_Analytical_Framing/05_Sources_And_Derivation_Convention.html.

@misc{wtpp_2026_103_sources_and_derivation_convention,
  author    = {Robertson, Jason},
  title     = {Sources and Derivation Convention},
  year      = {2026},
  publisher = {We The People Platform},
  version   = {3.8.484},
  url       = {https://wethepeopleplatform.com/_web_html/05_Analytical_Framing/05_Sources_And_Derivation_Convention.html},
  note      = {Document 103 of 149}
}