Coverage is the supported share of a preregistered applicable outcome map — not the share of whatever data happened to be found. Outcome families, units, and orientation rules are defined before results are examined. This codebook is that layer: the scorecard reads its orientation rules for GII-D and its families for GII-V coverage.
Implemented in @gii/codebook · 38 seed families across six domains · rendered live from the library.
Orientation rules#
| Polarity | Meaning |
|---|---|
| positive_is_benefit | An increase in the measure is a preregistered benefit; oᵢ resolves to +1 when the effect increases it, −1 when it decreases it. |
| positive_is_burden | An increase is a preregistered burden; signs reversed. |
| ambiguous | Genuinely non-oriented: oᵢ = 0, and the indicator stays in the oriented set O under a validated ambiguity rule. |
| magnitude_only | No orientation rule: oᵢ = null, excluded from O (counts toward magnitude only). |
| target_band | Beneficial within a band; oᵢ is graded in [−1, 1] by movement toward or away from the band. |
How families resolve direction
Only families with a signed rule can push GII-D. Ambiguous families stay in the oriented set at zero; magnitude-only families leave it entirely.
- ambiguous (o = 0)1334%
- benefit ↑1232%
- burden ↑1026%
- target band25%
- magnitude only (o = null)13%
38 families in total · The rules
Families per impact domain
A domain with no family is a domain the index cannot yet describe, so the flat profile here is the design target, not an accident.
38 families in total · Browse the codebook
Outcome families#
People & households
| Family | Measure · unit | Orientation | Sources |
|---|---|---|---|
| household_income | Real household income USD (constant) | benefit ↑ | Census Bureau APIs, BLS, IRS Statistics of Income |
| employment | Employment rate percent | benefit ↑ | BLS |
| consumer_prices | Consumer price inflation percent YoY | target band band withheld | BLS |
| health_coverage | Insurance coverage rate percent | benefit ↑ | CMS data, Census Bureau APIs |
| administrative_burden | Time and paperwork burden hours per person | burden ↑ | Agency administrative systems |
| service_utilization | Program/service utilization count or rate | ambiguous (o = 0) | Agency administrative systems, CMS data |
| poverty | Poverty rate (official and supplemental) percent | burden ↑ | Census Bureau APIs |
| housing_cost_burden | Share of households paying >30% of income for housing percent | burden ↑ | HUD USER API, Census Bureau APIs |
| energy_costs | Residential energy prices and expenditures USD or index | burden ↑ | EIA API, BLS |
| ui_claims | Unemployment-insurance initial claims count | burden ↑ | DOL API, BLS API |
U.S. economy
| Family | Measure · unit | Orientation | Sources |
|---|---|---|---|
| real_output | Real GDP USD (chained) | benefit ↑ | BEA API |
| inflation | Inflation rate percent YoY | target band band withheld | BLS API, Federal Reserve data |
| policy_rate | Policy interest rate percent | ambiguous (o = 0) | Federal Reserve data |
| public_deficit | Federal deficit USD or percent of GDP | ambiguous (o = 0) | Treasury Fiscal Data, CBO |
| employment | Nonfarm employment jobs | benefit ↑ | BLS API |
| labor_force_participation | Labor-force participation rate percent | benefit ↑ | BLS API, FRED API |
| business_investment | Real private nonresidential fixed investment USD (chained) | benefit ↑ | BEA API, FRED API |
| energy_prices | Wholesale and retail energy prices USD or index | ambiguous (o = 0) | EIA API |
Public entities
| Family | Measure · unit | Orientation | Sources |
|---|---|---|---|
| federal_awards | Federal awards received USD | ambiguous (o = 0) | USAspending API, Agency grants data |
| administrative_cost | Implementation and administrative cost USD | burden ↑ | Treasury, Census of Governments |
| service_capacity | Staffing and service capacity FTE or capacity index | benefit ↑ | Census of Governments, OpenFEMA |
| disaster_assistance | Federal disaster assistance obligations USD | ambiguous (o = 0) | OpenFEMA API, USAspending API |
Nonprofits & civil society
| Family | Measure · unit | Orientation | Sources |
|---|---|---|---|
| grant_revenue | Grant and contract revenue USD | benefit ↑ | IRS Form 990 data, USAspending |
| service_demand | Demand for services count | ambiguous (o = 0) | IRS Form 990 data, Agency grant systems |
| compliance_burden | Compliance and reporting burden hours or USD | burden ↑ | Agency grant systems |
| employment | Nonprofit-sector employment jobs | benefit ↑ | BLS establishment data, IRS Form 990 data |
International trade
| Family | Measure · unit | Orientation | Sources |
|---|---|---|---|
| imports | Import value USD | ambiguous (o = 0) | Census international trade data, USITC DataWeb |
| exports | Export value USD | ambiguous (o = 0) | Census international trade data, BEA international accounts |
| tariff_incidence | Tariff incidence / customs cost borne USD | burden ↑ | USITC DataWeb, Agency trade records |
| trade_balance | Trade balance USD | ambiguous (o = 0) | Census international trade data, BEA international accounts |
| import_prices | Import price index index | burden ↑ | BLS API, USITC DataWeb |
Private enterprise
| Family | Measure · unit | Orientation | Sources |
|---|---|---|---|
| firm_revenue | Firm revenue USD | benefit ↑ | Census business data, SEC EDGAR APIs |
| compliance_cost | Regulatory compliance cost USD or hours | burden ↑ | Regulators, SBA, Census business data |
| market_concentration | Market concentration (HHI) HHI | ambiguous (o = 0) | Census business data, SEC EDGAR APIs |
| firm_dynamics | Establishment entry and exit count | ambiguous (o = 0) | Census business data, BLS establishment data |
| enforcement_volume | Enforcement actions count | ambiguous (o = 0) | Regulators |
| business_applications | New business applications count | benefit ↑ | Census business data |
| market_response | Bundled entry, exit, investment, and vendor-demand response count or USD | magnitude only (o = null) | Census business data, SEC EDGAR APIs |
Magnitude anchors#
Each indicator’s severity, reach, persistence, and spillover is scored 0–100 against published anchors. Three of the four scales are quantified: each anchor carries a numeric band edge in a declared metric, so an estimate can be anchored deterministically by interpolation rather than by adjective. Spillover deliberately is not — propagation has no natural continuous metric, and quantifying it would be false precision. That one component stays structured judgment, which is exactly what the calibration run measures the reliability of.
The anchor scales themselves are not published. They are the conversion from a measured quantity into a score — the preregistered judgment that makes this an index rather than a collection of statistics — and publishing them alongside the family list would amount to publishing the instrument. What is published is the property that matters to a reader: the same measured input produces the same score every time, and the reliability of the judgment components is measured and reported rather than asserted.
Controlled missingness#
A blank value, a real zero, and the four analytic null-reasons are different states. A required field may be null only with a controlled reason, and missing weight is never silently redistributed.
| State | Meaning | Valid null reason? |
|---|---|---|
| value | A real observed value, including a genuine measured zero. Not missing. | no |
| zero | A measured value that happens to be zero — distinct from absent. Not a missingness reason. | no |
| not_applicable | The measure does not apply to this event, population, or domain. | yes |
| legally_unavailable | Data exists but cannot be used under law, license, or privacy rules. | yes |
| not_yet_observed | Exposure has not produced observable outcomes for this horizon yet. | yes |
| not_estimable | No credible counterfactual or method can currently estimate the effect. | yes |