How CadNews rates and compares news outlets

CadNews does not invent its own bias or reliability scores. Each outlet's political lean and factual-reporting rating comes from AllSides and Media Bias/Fact Check, and every outlet page links to the rater's own page for that outlet so you can check it.

• When both raters cover an outlet, we show both. For sorting and the left/right balance bar we use AllSides first, then Media Bias/Fact Check.
• When the raters disagree, both ratings stay visible. We never hide a rating that differs.
• When no rater has a page for an outlet, it is shown as unrated. Politics and world stories are built only from rated outlets.
• A sister outlet's rating (for example Yahoo News for Yahoo Sports) is shown only as labeled context, never as the outlet's own rating.

"Left-leaning" means outlets rated Left or Lean Left (Left-Center); "right-leaning" means Right or Lean Right (Right-Center); Center and Least Biased outlets count as center. When a count includes only outlets rated Left or Right, we say "most left-leaning" or "most right-leaning".

1. We read the public news feeds of about 150 outlets across the spectrum, including national, wire, local nonprofit, business, tech, sports and crypto outlets, several times an hour.
2. Articles about the same event are grouped into one story. Each headline links to the original article on the outlet's site.
3. For stories covered by several outlets, an AI model lists the facts the outlets agree on and summarizes how left- and right-leaning coverage framed the story differently. Outlet headlines are quoted as published.
4. Opinion pieces are excluded from outlet statistics.

Story pages appear in search engines only when at least 3 outlets covered the story and the comparison is complete. Thinner stories stay on the site but are marked noindex.

Outlet pages also show CadNews's own measurements from the last 30 days: how many stories included the outlet, which topics it covered most, and how often it covered the same stories as outlets on the other side.

For analyzed stories, the AI classifies how each outlet's article was framed, based on emphasis, word choice and what it includes or leaves out, and flags loaded phrases. Topic outlets use a lens that fits the topic: whether the core story was confirmed, industry versus skeptic perspectives, or how health and science evidence was described.

• Figures appear only after at least 5 analyzed stories; percentages need at least 10 analyzed articles.
• Articles with unclear or missing assessments are left out of the percentages.
• No figure is inferred from an outlet's reputation. Each one comes from articles we actually compared.

These are automated estimates, not verdicts. "Confirmed" describes the core story after comparing coverage, not an outlet's overall accuracy.

• We only see what outlets publish in their public feeds, so some articles are missed.
• AI summaries can be wrong. Every claim links back to the outlets' own articles so you can check.
• Outside ratings change over time. We follow the raters' published pages and update when they do.

See a mistake? Tell us and we'll correct it.