Methodology
How we cut through the noise.
Every story published on Gen Us goes through a rigorous editorial process. Here's what we look for and how we score it.
Propaganda Scoring
Every external article we analyze receives a Propaganda Score from 0-100. The score measures how many manipulation techniques are present and how aggressively they're used. Learn about all 12 propaganda techniques we track →
- •Loaded Language — emotionally manipulative words designed to trigger rather than inform
- •Ownership & Funding — who owns the outlet, their corporate parent, shareholders, political ties
- •Narrative Framing — what the story emphasizes, what it omits, who the framing serves
- •Source Diversity — single-perspective reporting vs. multiple viewpoints
- •Who Benefits — which entities benefit from the story's angle
- •Missing Context — relevant information the original article left out
Fact-Checking
Major claims in every story are verified against multiple independent sources. Each claim gets one of three statuses:
- •Verified — supported by multiple independent sources or public records
- •Unverified — cannot be independently confirmed, but not contradicted by evidence
- •Disputed — contradicted by credible sources or contains factual errors
Stories that fail fact-checking are not published. Stories with unverified claims are published with caveats.
Ownership Mapping
For every story we analyze, we trace the ownership chain of the publishing outlet:
- •Corporate parent company and subsidiary relationships
- •Major shareholders and board members with political ties
- •Known funding sources and advertising relationships
- •Cross-ownership with other media properties
Quality Standards
Every story is scored across five dimensions before publication. Stories that don't meet our threshold are rejected.
- •Factual Accuracy — 35% weight. Are claims properly sourced and verified?
- •Writing Quality — 25% weight. Is the prose clear, engaging, and professional?
- •Value Add — 20% weight. Does this add context beyond the original source?
- •Ethical Standards — 10% weight. Is the reporting fair and responsible?
- •Readability — 10% weight. Is it accessible to a general audience?
Propaganda Score Scale
Straightforward reporting with little detectable spin
Minor framing issues or slight omissions
Notable loaded language, missing context, or ownership bias
Significant propaganda techniques across multiple dimensions
Heavy manipulation — loaded language, omissions, and clear beneficiary alignment
Most mainstream outlet articles score between 20-60. The score reflects propaganda technique density, not political alignment — both left-leaning and right-leaning outlets can score high or low.
Source Monitoring
We monitor 39 outlets across the political spectrum. We apply the same methodology and scoring to all sources, regardless of their political leaning. Want to sharpen your own skills? Read our Media Literacy Guide.
Wire: Reuters, AP
Western Left: CNN, Guardian, BBC, NPR, MSNBC
Western Right: Fox News, NY Post, Daily Mail
Global South: Al Jazeera, SCMP, Middle East Eye, Deutsche Welle, France 24, Times of India, Japan Times, The Africa Report, TRT World
Investigative: The Intercept, ProPublica, Democracy Now, OCCRP, Bellingcat
Alternative: Grayzone, MintPress, ZeroHedge, Common Dreams, Jacobin, Mother Jones, Reason
Tech: The Verge, Ars Technica, TechCrunch, Wired, 404 Media
Finance: CoinDesk, Financial Times
Career pro-Israel PAC totals
When we report a member of Congress's career pro-Israel PAC total, we aggregate disclosed contributions from the universe of pro-Israel PACs tracked by OpenSecrets' “Pro-Israel” industry classification, plus the named AIPAC-aligned PACs we monitor directly (AIPAC PAC, United Democracy Project, NORPAC, Democratic Majority for Israel, and the Republican Jewish Coalition Victory Fund). Figures cover the member's full career through the most recent FEC filing cycle.
For lobbyist pages, “disclosed lobbying” figures are sourced from Senate Lobbying Disclosure Act (LD-2) filings, accessed via the Senate Office of Public Records and aggregated by OpenSecrets. Where a lobbyist's firm-level total is cited, we use the firm's aggregate federal lobbying revenue for the most recent complete reporting year; individual-lobbyist registrations within that firm are listed in the LD-2 filings themselves.
We do not estimate or extrapolate. If a figure is not present in a disclosed filing or a reputable secondary source (NYT, Washington Post, Politico, Reuters, AP, OpenSecrets), the field is left blank. Career totals are restated whenever the FEC or LDA databases are updated.
PAC spending
On PAC profile pages, the headline “recent spending” figure combines three FEC-reported components: (1) Schedule E — independent expenditures (ad buys, mailers, digital spending in support of or opposition to a specific federal candidate); (2) Schedule A receipts to candidate committees from the PAC, when applicable; and (3) inter-PAC transfers that move money from one PAC in the network to another. We follow the same convention used by the FEC and OpenSecrets when summing committee activity across an election cycle.
There is a known gap between “reported” spending — the figure that appears in our headline while a cycle is active — and the final reconciled number after the FEC completes its post-cycle audit and any amended filings. Amendments can move the total in either direction. We update PAC totals on a daily revalidation cycle as the FEC API publishes new filings, and we cite the FEC committee ID alongside each profile so readers can independently audit our number.
Shell PACs — pop-up committees created for a single race — are flagged in our data and on the PAC's profile. Their spending counts toward the headline figure of the parent or beneficiary committee only when an inter-committee transfer is documented in Schedule B disbursements.
Roll-call votes
Roll-call vote tallies (Yea / Nay) on bill pages are sourced directly from the official records published by Congress: Senate roll-call votes are pulled from the Senate's public XML feed (senate.gov/legislative/LIS/roll_call_lists), and House roll-call votes from the House Clerk's CSV/XML records (clerk.house.gov/Votes). Where applicable, we cross-reference Congress.gov, the canonical legislative record maintained by the Library of Congress.
Member names from the official records are matched to our internal politician registry using normalized name comparison: a nickname map (e.g., “Bob” → “Robert”), hyphen and apostrophe normalization, and state-disambiguation for shared surnames. Members listed as “Not Voting”, “Present”, or “Excused” in the official record are surfaced in the full roll call but do not factor into the Yea/Nay headline.
The lobby-money columns shown alongside each vote use the same career pro-Israel PAC totals described under Career pro-Israel PAC totals. We do not infer causation from the correlation between a vote and a member's career lobby total; we surface both numbers and let readers draw their own conclusions.
Donor totals
On donor profile pages, the “total reported” figure represents the sum of disclosed federal political giving by the individual (and, where the donor operates through a family LLC or family foundation, the giving attributable to those vehicles as reported in FEC filings). Sources are the FEC's individual contributor database, OpenSecrets' donor-lookup tool, and — for major donor profiles — Forbes' political-giving estimates where they are corroborated by FEC filings.
We distinguish between hard-money contributions to candidate committees (capped at the individual contribution limit per cycle), unlimited contributions to super PACs and 501(c)(4) dark-money organizations (the latter only where the donor or recipient has voluntarily disclosed the contribution or where the IRS Form 990 names the donor), and pass-through giving via joint fundraising committees. Each component is itemized on the donor's profile.
Career donor totals are restated when new FEC filings or amended returns are published. Notable single contributions are linked directly to the underlying FEC filing in the receipts section of every donor page so readers can verify any individual number.