Technical methodology · Expert-review edition

Household Resilience Scoring Methodology

A transparent index of household capability coverage—not a prediction of survival, damage, or uninterrupted self-sufficiency.

Model version: 2.0.0 · Document revision: October 23, 2026
Scope: full household assessment. Short Resilience Checks, demonstration profiles, ReadyNow, and projected roadmap gains are separate products or calculations.

Validation status: This is an initial, deterministic PREPHAUS policy model. Its weights, hazard margins, evidence discounts, coverage thresholds, and critical-gap ceiling have not been empirically calibrated or independently validated. Citing government preparedness guidance does not imply agency endorsement of the index.

1. What the index measures

The index compares available household capabilities with explicit requirements for a selected disruption horizon: 3, 7, 30, 180, or 365 days. It answers, “How much of this model’s preparedness requirement is covered?” It does not estimate expected loss, annual event probability, mortality, structural performance, or a guaranteed number of survivable days.

Each household receives a separate result for each horizon. Missing responses receive zero quantitative capability credit, which means not established, not proof that a capability is absent. Longer-horizon previews can therefore be incomplete when an assessment did not ask that horizon’s replenishment questions.

Thirty-second explanation: “PREPHAUS compares household-reported capabilities with duration-specific preparedness requirements across seven weighted areas. It adjusts selected requirements for regional hazard context, discounts unverified evidence, and prevents designated critical gaps from being hidden by a high average. The result is an explainable preparedness index, not a certified safety rating.”

2. Factors and headline weights

Household water days; tested replenishment-source response for 180- and 365-day horizons.Complete household days of food.Explicitly reported cellular alert/contact capability and one non-cellular channel.Complete plan, practiced drill, and non-elevator egress for apartments.Shortest essential medication or medical/first-aid supply duration.Sanitation continuity; independent wildfire, hurricane-opening, and tornado-shelter responses when applicable.
PillarWeightScored requirements
Backup Power20%Critical-load runtime; separate cooling-zone runtime when heat applies.
Water18%
Food14%
Communications12%
Evacuation12%
Medical & First Aid12%
Storm Protection & Shelter12%

Requirements within a pillar share its weight equally. For example, when cooling applies, critical-load runtime and cooling runtime each receive 10 percentage points of the 20% power weight. Adding a conditional requirement never increases the headline weight. Wildfire mitigation remains inside the same 12% protection pillar; it does not use an unmatched alternate weight.

PREPHAUS policy choices

The seven weights and equal allocation within a pillar are design assumptions. They are not FEMA, CDC, actuarial, or engineering coefficients.

3. Duration requirements and hazard margins

Base requirement3 days7 days30 days180 days365 days
Critical-load runtime, hours24562401,4402,920
Water, food, medical, sanitation: household days3730180365
Communication channels11222
Complete evacuation plan; practiced drill1 each1 each1 each1 each1 each
Tested water replenishment sourceNot requiredNot requiredNot required11
  • Heat or cold: multiply critical-load power hours by 1.25. The eight-hours/day base is a PREPHAUS planning assumption, not continuous whole-home operation.
  • Heat or flood: multiply water days by 1.20, rounded to one decimal place. CDC supports extra water for hot climates; the exact 20% factor and flood application are PREPHAUS policy.
  • Heat: separately require eight cooling-zone hours/day. Report runtime remaining after allocating shared energy to other critical loads. This is not a thermodynamic or indoor-temperature safety model.
  • Apartment: require a separate non-elevator egress plan.
  • Wildfire: require reported first-five-feet combustible clearance.
  • Hurricane: require reported rated opening protection.
  • Tornado: require reported accessible designated shelter/safe-room capability.

Matching hazards do not compound the same power or water margin. A tornado-shelter response cannot satisfy wildfire mitigation, and defensible space cannot satisfy hurricane-opening protection. Flood context currently changes water demand but does not establish flood-safe shelter, elevation, insurance, or floodproofing.

Long horizons are requirements, not certified supply: Version 2 uses all 180 or 365 days instead of the old 60/90-day proxy. A “yes” to a water-source question does not create any additional water days. Yield, potability, seasonality, energy, access, food production, fuel replenishment, and medical supply chains remain unverified. Do not interpret the result as a prescription to stockpile a year of medication; continuity planning must follow clinician/pharmacist guidance.

4. Translating household responses into capability

  • Water: use reported household days. When stored gallons are also answered, the credited starting quantity is the lower of reported days and gallons ÷ household size, using a one-gallon/person/day planning floor. This conservative cross-check never inflates supply. It can understate a replenishable system and does not establish treatment effectiveness or higher individual needs.
  • Food, medical, sanitation: use nonnegative reported household days. The model does not independently measure calories, dietary suitability, care skills, medication safety, or sanitation-system performance.
  • Power and cooling: keep runtimes separate. A reported load test does not add 10% runtime or confer independent verification. Fuel, safe generator placement, battery degradation, and actual simultaneous load schedules are not simulated.
  • Communications: count zero, one, or two explicitly reported paths. No phone channel is assumed from a missing answer. Broadcast warning reception is useful but is not equivalent to two-way messaging.
  • Plan, drill, egress, protection, shelter, source: yes = 1 and no/missing = 0 before evidence adjustment. A drill contributes separately; it does not multiply plan capacity. Missing or negative plan response prevents drill credit.

A household’s reports are not automatically credible because they are numeric. The capability is still self-reported unless a documented or independently verified evidence status has actually been established.

5. Evidence rules

StatusUsable capabilityMeaning
Reported / no qualifying evidenceReported capability × 0.90Normal onboarding status; no independent confirmation.
DocumentedReported capability × 0.95Documentation status exists; not equivalent to independent inspection.
Verified with a nonnegative numeric verified valueVerified valueThe verified quantity takes precedence, including verified zero.

The 90% and 95% factors are conservative PREPHAUS adjustments, not statistical confidence levels. A fully self-reported yes/no requirement normally scores 90, not 100. A test/drill response is not an upload, inspection, or verification event. This revision does not establish an accredited inspector network or validate the provenance of every caller-supplied evidence status; evidence authorization requires further review before high-assurance claims.

6. Exact calculation and rubric

usable_i = evidence-adjusted capability in the requirement’s exact unit
score_i  = clamp(round(100 × usable_i / max(required_i, 0.01)), 0, 100)
weight_i = headline pillar weight / applicable requirements in that pillar
average  = round(sum(score_i × weight_i) / sum(weight_i))

If ANY designated critical requirement has score_i < 50:
  overall = min(average, 49)
Otherwise:
  overall = average

The cap evaluates rounded requirement scores. At exactly 50, that requirement does not activate the cap. This is an explicit policy boundary, not a biological safety threshold. Pillar display scores are the average of their applicable requirement scores; the overall critical-gap ceiling does not rewrite those individual scores.

Critical requirements: water days, food days, medical days, sanitation days, complete evacuation plan; plus apartment egress, heat cooling, tornado shelter, and long-horizon water replenishment when applicable. Power runtime, communications, drills, combustible clearance, and opening protection remain scored but do not independently trigger this ceiling.

Score bandPillar labelInterpretation
75–100StrongHigher modeled coverage, not a safety certification.
50–74DevelopingMaterial improvement remains.
0–49Critical GapLow coverage and/or the overall critical-gap ceiling.

7. Worked example

Consider a three-day, non-heat household with exactly three days of water, food, medical supplies, and sanitation; 24 hours of critical-load power; one explicitly reported phone channel; and a complete, practiced evacuation plan. All evidence is self-reported.

  1. Each of those requirements scores 90 after the 0.90 adjustment.
  2. Each pillar therefore scores 90, and the weighted average is 90.
  3. No designated critical requirement is below 50, so the final index is 90.

If essential medical continuity is instead zero, its 12% contribution drops from 10.8 to zero. The rounded weighted average is 79, but the critical medical requirement activates the ceiling. The final index is 49, with the medical shortfall explicitly shown. A favorable average can no longer conceal that gap.

For heat exposure, the same three days of reported water become 2.7 usable days against a 3.6-day requirement: round(2.7 / 3.6 × 100) = 75. Separate cooling and increased power requirements also apply.

8. Data provenance: what is actually queried

InputImplemented source / behaviorBoundary
Household capabilitiesAssessment responses and associated evidence records.Usually self-reported. No automatic inference from products purchased or money spent.
Address and geographic contextU.S. Census geocoder.A failed resolution now stops analysis. No keyword-selected demonstration location is substituted.
Weather point/zone and current alertsLive National Weather Service point metadata and CAP alerts.Live alerts are not a climatological loss model and do not become household capabilities.
Earthquake contextUSGS recent-event query and a PREPHAUS normalization heuristic.Recent events are not a long-term seismic hazard estimate or structural assessment. Failure uses a labeled reference fallback.
Wildfire, heat, flood, tornado, air-quality baseline indicesPREPHAUS state/regional numerical rules with background links to relevant agencies.Not fresh USFS, FEMA NFHL, SPC climatology, or AirNow dataset measurements; no parcel precision or measured dataset vintage is claimed.
USGS water-monitoring providerNot connected to baseline scoring.Listed as unavailable, not as a live supplied observation.

Onboarding selects regional hazard tags with normalized indices ≥50, or the top-ranked tag if none qualify. That threshold is PREPHAUS policy. The regional index selects questions and margins; it is not averaged into the household capability score. Some hazards, including hurricanes and cold, are supported by the rubric but are not comprehensively identified by the current baseline regional adapter.

New real-household results no longer inherit numeric hazard cards from demonstration properties. This does not replace the regional adapter with authoritative parcel data; a separate data-integration project is still necessary.

9. Guidance citations and policy/source separation

The following pages were inspected for this review. They support preparedness practices, not PREPHAUS’s mathematical coefficients.

  1. CDC — How to Create an Emergency Water Supply. Supports at least one gallon/person/day for three days, trying for two weeks where possible, and additional water for hot climates and certain needs. Does not establish PREPHAUS’s 18% water weight, 20% margin, or year-long target.
  2. Ready.gov — Build A Kit. Supports water, food, radios, first aid, medication planning, and sanitation supplies. Does not validate the seven weights or evidence discounts.
  3. Ready.gov — Evacuation. Supports destinations, alternative routes, transport, dependents, alerts, and household contact plans. Does not establish a numerical plan score.
  4. Ready.gov — Power Outages. Supports critical-device planning, safe backup power, medication continuity and powered community locations during extreme temperatures. Does not establish eight operating hours/day or certify a safe cooling runtime.

Other reference destinations: Census geocoding, NWS API documentation, USGS Earthquake Hazards Program, FEMA flood maps, and Wildfire Risk to Communities. A link alone is not evidence that the corresponding dataset was ingested.

10. What changed from version 1.0.0

  • Enforced the critical-gap ceiling instead of relying only on a weighted average.
  • Connected sanitation and apartment egress to explicit requirements.
  • Separated critical-load power, cooling, and independent hazard mitigations; removed the load-test runtime bonus and fractional plan boost.
  • Removed assumed communication capability from unanswered questions.
  • Used complete 180-/365-day requirements, with an additional explicit water-source requirement, instead of 60-/90-day proxies.
  • Kept protection weights consistent across wildfire and other contexts.
  • Made verified zero authoritative rather than falling back to a positive report.
  • Removed substitute geocoding locations and demo hazard inheritance from new household output; labeled regional assumptions honestly.

Historical persisted results are not rewritten. Legacy model snapshots and recorded contributions without critical metadata retain their original scoring semantics. A fresh assessment under a different version is not proof of real-world improvement or deterioration. Version-specific comparisons and new assessment answers are required.

11. Boundaries with other PREPHAUS features

  • ReadyNow: a separate operational layer applies current-condition deductions and maintenance-confidence context; it does not rewrite the baseline assessment.
  • Short Resilience Check: its own categorical short-check methodology; not upgraded to a full assessment by this revision.
  • Demonstrations: labeled examples with separate simplified calculations, not household evidence.
  • Guide scenario calculations: reuse recorded requirement contributions and critical metadata. Product deltas must have matching capability units; a purchase is not verified installation or achieved preparedness.
  • Roadmap projections: estimated gains are not achieved score changes, and must not be treated as audited evidence of completed capability.

12. Known limitations and appropriate use

Appropriate use: explain gaps, organize preparedness discussions, compare requirement coverage within the same model version and horizon, and identify questions needing professional assessment.

Inappropriate use: insurance underwriting, life-safety certification, structural certification, parcel risk certification, predicted casualty reduction, guaranteed restoration time, or medical stockpiling instructions.

  • No empirical calibration, outcome study, inter-rater reliability study, or uncertainty interval.
  • Limited personalization for disability, age, pregnancy, pets, special diets, medical devices, evacuation accessibility and no-prescription households.
  • Regional hazard rules are coarse, incomplete, and can select inappropriate or miss relevant questions.
  • No full renewable-yield, fuel, potable-water, food-production, energy-load, cold-chain, or dependency simulation.
  • Self-report and evidence-status authorization require further assurance controls; stale records are not automatically discounted in this baseline.
  • The binary mitigation questions are proxies, not inspections of fire hardening, flood safety, wind resistance or shelter engineering ratings.
  • Medical continuity remains a crude duration proxy; a household with no prescription need lacks an explicit not-applicable pathway.

13. Expert-review and validation agenda

  1. Content validity: ask emergency management, public health, accessibility, utility/power, fire, and structural specialists to review each requirement and missing domain.
  2. Policy calibration: review weights, equal within-pillar allocation, eight-hour power/cooling assumptions, margins, discounts, score bands, and the 49-point ceiling.
  3. Source-backed exposure: replace state heuristics with appropriate authoritative geography-specific datasets; record version, retrieval time, resolution, failure behavior and permitted use.
  4. Evidence governance: define who may assign documented/verified status, permissible evidence, inspection scope, freshness and revocation.
  5. Scenario coverage: test evacuation versus shelter-in-place, apartment accessibility, medically dependent households, hot/cold climates, outages and compound hazards.
  6. Validation study: compare scoring with independent assessor judgments, test repeatability and sensitivity, and only then investigate association with observed preparedness outcomes.
  7. Release governance: publish a versioned change log, keep historical evidence immutable, communicate score discontinuities and monitor errors.

Questions for the expert: Which capabilities are truly non-compensable? Which requirements should be conditional or not applicable? What evidence establishes operational capability? Which duration assumptions are locally defensible? What claim would you permit after reviewing this model—and what evidence would you require first?

This document describes the built-in full-assessment model, not an independently certified standard. Administrative custom models can override configured requirements and need their own reviewed methodology. Use your browser’s Print → Save as PDF to share a print-ready copy.