Last reviewed: April 23, 2026. Data itself refreshes on the cadence described at the bottom of this page.
FederalVoice uses a local-first data pipeline built on public-source records to assemble profile summaries, scorecards, and campaign finance visuals. Everything below is derived from data the U.S. government publishes — no private scoring models, no paywalled sources.
Core Inputs
- Congress API v3 — member records, roll-call metadata, bill context, and vote history.
- FEC Open API — candidate finance totals, outside spending, and contribution aggregates.
- House and Senate financial disclosure portals — annual net-worth range disclosures.
- Local profile metadata — role, office, contact fields, and editorial context stored per member in
content/bio/*.md.
Scorecard Methods
Ideology index (0–100 scale)
How consistently a member votes with the left or right of the chamber, derived from their roll-call voting pattern in the active Congress window. Lower values = more left-leaning; higher values = more right-leaning. Middle ranges indicate cross-party voting. This is not a policy-position score; it’s a behavioral measure of how a member has actually voted.
Attendance
Percentage of tracked roll-call votes the member cast versus missed. Missed votes include absences, not-voting announcements, and present-but-not-voting events.
Party loyalty
How often the member voted with their party’s majority on Yea/Nay positions. A 90% loyalty member breaks with their party on roughly 1 in 10 contested votes.
Recent votes displayed on profile pages are a live feed, not part of the score baseline — scoring uses the broader Congress-window sample.
Finance Methods
- Receipts, cash on hand, and individual share come directly from FEC candidate totals.
- Outside spending combines support and oppose totals from schedule E summaries.
- Top occupation/employer values come from FEC schedule A aggregation endpoints (donors self-report these fields).
- Trend charts use available multi-cycle candidate totals.
Net-Worth Estimates
Members file annual public disclosures that list each asset and liability inside a range — e.g., “$1,000,001–$5,000,000” — not an exact dollar figure. FederalVoice:
- Parses every range in each filing.
- Sums the low end of every asset range and subtracts the high end of every liability range to get a conservative floor.
- Sums the high end of every asset range minus the low end of every liability range to get a ceiling.
- Reports the midpoint as the estimate, with the floor–ceiling band shown alongside.
Because disclosures use ranges, the estimate is inherently approximate. It will never match an exact bank-statement figure and is not intended to.
Transparency Principles
- Neutral, descriptive language over partisan framing.
- Timestamped refresh metadata on generated snapshot records.
- Primary-source links surfaced on profile and analytics pages.
- Known data gaps are shown explicitly rather than silently filled.
Data Cadence
- Current-member activity snapshots refresh daily via admin scripts.
- Campaign finance enrichment runs hourly for current members with cycle-aware freshness checks.
- Net-worth dataset refreshes as new disclosures are filed.
- Historical members may rely on curated local overrides when federal APIs are incomplete.