How the Compliance Risk Index is built.
One canonical page for every source, computation, and disclosed limitation behind the Compliance Risk Index โ so a journalist, researcher, or compliance officer can verify any number on that page without asking us first.
The five real sources
Update cadence
OSHA is the only one of the five on an automated schedule today: a real cron job re-aggregates the already-downloaded extract every 3 hours, and pulls a fresh copy of the underlying DOL dataset quarterly (1st of January, April, July, and October). EPA RCRA, FMCSA, CMS nursing-home, and BLS OEWS are refreshed by a real, disclosed manual pull as each source publishes a new extract โ not yet on a fixed automated cadence. Every dataset section on the Risk Index page shows the real date it was last refreshed, computed from the actual ingestion timestamp, not a display placeholder.
How each metric is computed
- Volume โ a real count of citation, violation, or deficiency line-items under that CFR part / 29 CFR standard / BASIC category / deficiency category, within the dataset's real coverage window.
- Severity (OSHA) โ the real share of citations OSHA itself classified Serious, Willful, or Repeat (VIOL_TYPE S/W/R) among citations with a recorded VIOL_TYPE. OSHA's own classification, not a proxy we invented.
- Exposure (RCRA, OSHA) โ the real share of violations/citations at facilities classified Small or Very-Small Quantity Generator (RCRA) or reported under 100 employees (OSHA) โ a facility-size signal, not an employee-headcount claim.
- Avg. penalty โ the real average penalty amount among cited facilities/establishments with a recorded enforcement action of that type โ a facility-level correlation, not a strict per-violation attribution (one facility can carry multiple violations under one penalty).
- OOS rate (FMCSA) โ the real share of inspections in that BASIC category that resulted in an out-of-service order, as recorded in FMCSA's own extract.
- Wage percentiles (BLS) โ real OEWS cross-sectional wage-distribution percentiles for the May 2025 reference period. The 10th percentile is a legitimate proxy for "entry-level pay" today; the 90th for "experienced/ceiling pay" today โ not a career-tenure or time-series measurement.
Known limitations
OSHA and RCRA enforcement coverage is federal only โ state-plan programs (Cal/OSHA, Washington, Oregon, and similar) run their own citation codes and aren't in these extracts. FMCSA's ingested extract carries no per-inspection state field, so that table is national-only regardless of the state filter elsewhere on the page. BLS withholds a state-level wage estimate whenever too few employers reported in that state; a missing state row means BLS itself suppressed the number, not that CertQuestUSA estimated around it. OSHA Safety & Health, HIPAA/Healthcare Compliance, and Workplace Compliance/HR exams mostly have no BLS ranking because they're training taken across many occupations rather than tied to one distinct SOC code. Every gap above is shown as "no CQU exam yet" or an absent row on the live page โ never filled in with an estimate.
Version history
Found an error?
Every number here traces to a named federal source with its real URL above โ if you find a mismatch, email us with what you found and we'll check it against the source and correct it fast.
