Methodology

What is SGO Data?

In June 2021, NHTSA issued Standing General Order 2021-01 (SGO), requiring manufacturers and operators of vehicles equipped with Automated Driving Systems (ADS) or Level 2 Advanced Driver Assistance Systems (ADAS) to report certain crashes. This was the first mandatory, standardized crash reporting requirement for AV-equipped vehicles in the United States.

ADS-equipped vehicles must report any crash within 24 hours. ADAS-equipped vehicles must report crashes involving a fatality, injury requiring hospitalization, airbag deployment, pedestrian/cyclist involvement, or a vehicle tow-away.

Data Processing Pipeline

We process the raw SGO data as follows:

1. Ingestion: Raw reports are downloaded from NHTSA's public SGO dataset as structured CSV/JSON files.
2. Parsing: Each report is parsed into structured fields including date, location, vehicle information, system type, and severity.
3. Deduplication: Reports are deduplicated by Report ID, keeping only the highest version number. When a manufacturer submits an updated report for an existing incident (e.g., adding injury details discovered later), the newer version supersedes the original. This means our incident count reflects unique crashes, not total filings.
4. Classification: Incidents are classified as ADS or ADAS based on the reporting manufacturer's designation in the SGO filing.
5. Severity assignment: Each incident is classified into one of three severity levels (see below).
6. Aggregation: Statistics are aggregated by manufacturer, vehicle model, state, and time period to power the site's various views.

Crash Severity Classification

We classify each incident into one of three severity levels based on fields in the SGO report:

ADS vs ADAS Classification

The distinction between ADS (Automated Driving System) and ADAS (Advanced Driver Assistance System) is critical for understanding the data:

• ADS (SAE Level 3-5): The vehicle's automated system handles all driving tasks. The human is not expected to be actively driving. Examples: Waymo robotaxis, Cruise driverless vehicles, Zoox shuttles.
• ADAS (SAE Level 2): The system assists the driver but requires continuous human supervision. Examples: Tesla Autopilot/FSD, GM Super Cruise, Ford BlueCruise.

We use the manufacturer's own classification from their SGO filing. We do not independently verify whether a system qualifies as ADS or ADAS. Note that Tesla's "Full Self-Driving" is classified as ADAS (Level 2) in SGO filings, with the recent exception of a small number of ADS-designated reports.

Deduplication Details

NHTSA's SGO system assigns each incident a unique Report ID(e.g., "30530-14071"). Manufacturers may submit multiple versions of the same report as more information becomes available — for example, when an initially non-fatal crash results in a later death, or when police report details are updated.

Our pipeline keeps only the highest version of each Report ID. This ensures each physical crash is counted once with the most complete information available.

Manufacturer Rankings

Our manufacturer rankingsare based on total SGO incident count. We rank by raw count by default because it's the most transparent metric available. However, we strongly caution against interpreting raw rankings as safety rankings:

• A manufacturer with more vehicles on the road will naturally report more incidents
• ADS operators report all crashes (even minor ones), while ADAS manufacturers only report serious crashes — this inflates ADS counts
• Without miles-driven normalization, raw counts are misleading for cross-manufacturer safety comparisons

We provide fatality counts, injury counts, ADS/ADAS splits, and per-vehicle breakdowns to allow more nuanced analysis.

Data Freshness

We update our database on a monthly cadence:

• SGO data: Pulled monthly from NHTSA's published dataset. NHTSA typically updates the public dataset every 1-2 months.
• Complaints: Pulled via NHTSA's API, filtered for AV/ADAS-related keywords.
• Recalls: Pulled via NHTSA's Recalls API, filtered for autonomous/ADAS-related components.
• Investigations: Pulled from NHTSA's ODI database.
• Safety ratings: Pulled from NHTSA's NCAP Safety Ratings API for vehicles in our database.

How We Score Safety

Our manufacturer safety scorecards combine multiple dimensions of data into a holistic picture. Here's how scoring works:

• Data freshness: All data is pulled from NHTSA on a monthly cadence. Scores reflect the most recent public dataset available.
• Incident classification: ADS (fully autonomous) and ADAS (driver-assist) incidents are tracked separately. Because ADS operators must report all crashes while ADAS manufacturers only report serious ones, we never combine raw counts across these categories for ranking purposes.
• Severity weighting: Fatal incidents carry the highest weight, followed by injury crashes, then property-damage-only incidents. This ensures that a single fatality is not treated equivalently to a minor fender-bender in aggregate scores.
• Recall & complaint signals: Active recalls, open investigations, and high complaint volumes are factored in as supplementary safety indicators alongside crash data.

We deliberately avoid producing a single "safety rank" number because, without miles-driven normalization, any such ranking would be misleading. Instead, we present multi-dimensional data and let researchers draw their own conclusions.

Known Limitations

• Reporting thresholds differ: ADS has a lower reporting threshold than ADAS, making raw count comparisons misleading. A Waymo robotaxi must report a minor parking lot bump; a Tesla with Autopilot does not.
• No exposure normalization: Without miles-driven data per manufacturer, we cannot calculate true crash rates. More vehicles = more reports, all else equal.
• Fleet size varies enormously: Tesla has millions of ADAS-equipped vehicles on the road; some ADS operators have hundreds. Raw counts reflect this disparity.
• Self-reported data: Manufacturers self-report incidents. Reporting quality, completeness, and timeliness may vary. There is no independent audit of compliance.
• Reporting lag: Incidents may take weeks or months to appear in the public dataset. Initial reports may lack details that are added in subsequent versions.
• CBI redaction: Some fields in SGO reports are redacted as Confidential Business Information (CBI). This may include software version, specific system details, and narrative descriptions.
• Not all AV crashes are included: Only crashes meeting SGO thresholds are reported. Minor fender-benders involving ADAS vehicles may not appear.
• Causation is not determined: SGO reports do not establish whether the ADS/ADAS caused the crash. The system may have been engaged but not at fault.

Interpretation Guidance

When using this data, remember: more incidents does not necessarily mean less safe. A manufacturer with more vehicles on the road will naturally have more reported incidents. Always consider fleet size, reporting thresholds, and miles driven when drawing conclusions. See our glossary for definitions of key terms.