RESEARCH STANDARD

From source record to repair decision.

Our methodology keeps applicability, evidence strength, freshness, uncertainty, and safety visible at every step.

Last reviewed August 25, 2026

1. Source hierarchy

We prefer primary, attributable records. No single source proves every issue, so each source is used only for the claim it can support.

  1. Safety authorities: recalls, investigations, and formal agency records.
  2. Manufacturer records: communications, bulletins, warranty extensions, parts, and repair procedures.
  3. Technical standards: diagnostic-code definitions and system conventions.
  4. Observed outcomes: documented repairs and owner complaints, used as signals with sampling limits.
  5. Secondary reporting: context and leads, not a substitute for the underlying record.

2. Vehicle applicability

A claim is mapped to the narrowest defensible application: make, model, generation, model year, engine, transmission, drivetrain, market, and build window. Unknown fields remain unknown; they are not filled by inference for SEO coverage.

A bulletin can describe a repair without proving failure rate. A complaint can show a reported symptom without proving cause. A recall applies only to the vehicles and remedy named in the campaign.

3. Evidence and confidence

Issue-family labels communicate evidence maturity, not a numeric probability that a specific vehicle will fail.

  • Established: consistent primary records and a stable application/remedy pattern.
  • Developing: credible signals exist, but scope, frequency, or remedy remains incomplete.
  • Preliminary: early signal suitable for monitoring, not a buying or repair conclusion.

4. Repair-cost ranges

Cost ranges separate diagnosis, parts, labor, and application assumptions. Production ranges will identify observation window, geography, parts grade, labor-rate basis, sample size where available, and last refresh.

Taxes, fees, rust, calibration, hidden damage, warranty, and unresolved causes can materially change a real estimate. Published ranges are not quotes and should not be represented to a shop as binding prices.

5. Freshness and revision chain

Records are designed to retain the source URL or identifier, retrieval date, source effective date, editorial review date, and successor or superseded revision. A freshness badge reports process state; it does not turn an old source into a current conclusion.

6. Safety and diagnostic limits

We distinguish “may continue with caution,” “schedule diagnosis,” and “stop and assess” guidance, but the observed condition controls. Online content cannot inspect the vehicle or clear it as safe.

Immediate warning patterns include loss of braking or steering, smoke or fire, a fuel leak, severe overheating, a red oil-pressure warning, violent control problems, or a flashing check-engine light with severe shaking. Move to safety when possible and use emergency, roadside, towing, or qualified repair help as appropriate.

7. Corrections and page eligibility

Material corrections should identify what changed and when. Public pages require a real user decision, sufficient evidence, a distinct application, and a review owner. Database combinations that would create thin or duplicative pages remain unpublished or noindexed.

See how these standards affect every vehicle record.

Read our methodology