Our Methodology

The numeric score (0–100) is always the honest, unmanipulated number. It is never curved. What the letter grade does is place that score in context: how does this representative compare to their peers in the same chamber?

House members are graded on a House curve. Senators are graded on a Senate curve. This means the same numeric score might earn different letter grades in each chamber, depending on the distribution. The numeric score tells you the absolute performance; the letter grade tells you the relative standing.

Absolute excellence thresholds: Regardless of the curve, exceptional scores earn top grades automatically. Very high scores always receive top-tier grades, ensuring that genuine excellence is recognized even if the entire chamber performs well.

Hard floors: Very low scores cannot receive passing grades regardless of where they fall on the curve. This ensures that poor absolute performance is never masked by a weak peer group.

Why a curve? Because an absolute grading scale would be misleading. If every member of Congress scored between 65 and 85, an absolute A–F scale would cluster everyone into two or three letter grades with no differentiation. The curve ensures the letter grade always communicates meaningful relative standing. The numeric score remains the honest, uncurved number.

Confidence indicator: Each grade includes a confidence level based on data volume — Low (under 3 months of data), Medium (3–9 months), or High (9+ months). New members receive grades with a Low confidence flag rather than being excluded. Missing data is scored as neutral, not penalized.

For each of the representative's stated or campaign positions across 24 policy topics, we identify votes on related bills. The score reflects the percentage of votes that align with stated positions.

Key principles:

• Inferred positions are excluded. We only measure against positions the representative has publicly stated or campaigned on. Contradicting a position we inferred from the voting record would be circular logic.
• Position specificity required. A broad claim like "I support veterans" is not specific enough to flag any veterans-related vote as a contradiction. The position must be specific enough to apply to the exact provision voted on.
• Omnibus bills treated carefully. Voting against a large spending package that funds something a member supports can reflect principled objection to other provisions — not hypocrisy.
• Procedural votes excluded. Motions to table, recommit, cloture, and similar procedural actions are never counted.
• Party-line skepticism. When nearly an entire party votes the same way, contradictions are flagged but weighted with additional context.
• Explained votes honored. If a representative publicly explains why they voted against their stated position, that vote is excluded from the penalty.
• Position changes respected. If a representative says "I changed my mind on this issue," we update the record. People are allowed to evolve — they just have to own it.

Transparency measures two things: does the representative take clear positions on major issues, and do they engage with media outside their ideological comfort zone?

Positions coverage: We track whether the representative has stated positions on major policy topics. Representatives with clear, public positions score higher. Empty or vague positions reduce the score.

Media transparency: We analyze the representative's media appearances across outlets classified by ideological lean. A Republican who appears only on conservative media, or a Democrat who appears only on progressive media, scores lower than one who engages across the spectrum. Appearing on outlets that challenge your positions is a positive signal — it demonstrates willingness to face scrutiny. Representatives who avoid challenging media environments entirely receive a reduced score on this sub-component.

This component exists because political science research is unambiguous: dehumanizing and inflammatory rhetoric from elected officials corrodes democratic norms, increases political violence, and harms democratic institutions — regardless of party. We track this not as a policy judgment but as a conduct standard.

A standardized keyword taxonomy applies to every representative regardless of party. Language is categorized by severity, with dehumanizing terms and incitement language weighted more heavily than common partisan rhetoric. A weighted rate is computed across all public posts, transcripts, and floor speeches.

Nonpartisan by design: The same words trigger the same penalty regardless of who says them. Dehumanizing language, incitement rhetoric, and personal attacks are flagged identically whether spoken by a Democrat or a Republican. The standard is conduct — not ideology.

Academic basis: This component is grounded in peer-reviewed research on the effects of elite rhetoric on democratic institutions, including work by the Carnegie Endowment for International Peace (2023), research on inflammatory rhetoric and political violence in Security Studies (Piazza, 2023), and the Political Hostility Scale framework published in Tandfonline (2025).

Are they focused on their district, or performing for national cameras?

Three sub-components:

• Donor geography: What percentage of campaign donors are from the representative's state or district? Higher in-district and in-state donation rates score better.
• Local social media focus: Does the representative's social media content address local and district-level issues, or is it predominantly national talking points?
• Town hall engagement: Does the representative hold public town halls where constituents can ask questions? More frequent, in-person town halls score higher. (See the Town Hall Accountability Grade section below for the separate, more detailed grading scale.)

Do their words match their other words? This component catches hypocrisy — specifically, representatives who publicly call for civility and then engage in the exact conduct they condemned.

Civility claims include explicit directives ("tone down the rhetoric"), self-imposed standards ("I believe in civil discourse"), and the implicit baseline accepted by taking office. When a representative makes such a claim and then violates it, that contradiction is flagged with a severity level based on how close in time the contradiction occurred to the original claim.

Note on the difference between Rhetorical Conduct and Rhetoric Consistency: Rhetorical Conduct measures behavior — how the representative actually speaks, regardless of what they've claimed. Rhetoric Consistency measures hypocrisy — the gap between what they say about civility and how they actually behave. A representative who is consistently inflammatory scores low on Conduct. A representative who claims to be civil while being inflammatory scores low on both.