Editorial Policy

Data first

Every page is built around a fixed set of figures from NHTSA and EPA: complaint counts by component, crashes, fires, injuries, deaths, recall campaigns and fuel economy. The tables and charts are generated directly from that data and are the primary content. Text exists to explain the numbers, never to replace them.

How the text is written

Short summaries, per-year notes, symptom descriptions and FAQ answers are drafted with a large language model. The model receives only the structured data and complaint narratives for that specific page and is instructed to use nothing else. An automated validator then checks every number in the draft against the source data and rejects drafts containing figures, prices, expert quotes or claims that are not in the input. Rejected drafts are regenerated or the page ships without text.

We deliberately do not publish:

• repair cost estimates (no reliable public source exists);
• invented expert opinions, testimonials or user reviews;
• verbatim complaint narratives, VINs, names or locations from NHTSA records;
• “reliability ratings” — our Complaint Index is a relative measure of owner complaints and recalls, as explained in the methodology.

Human review

The most-visited model pages are read and edited by a human editor. Edited text takes precedence over generated drafts and is not overwritten by later data refreshes.

Corrections

If you find an error, use the contact page. Data errors that trace back to NHTSA or EPA records are reported upstream; presentation errors are fixed in the next build, usually within a week.

Advertising

The site is funded by clearly labeled advertising and related-search units. Advertisers have no influence on verdicts, rankings or text. Ad units are never disguised as content or navigation.