TEA Scoring Methodology
This is the complete, public scoring methodology behind every TEA grade. Every criterion, every formula, every data source is published here. Anyone can audit it. We believe a grading system that asks politicians to be transparent must itself be transparent.
How TEA Grades Work
Every candidate is scored on three axes: Transparency, Efficiency, and Affordability. Each axis is scored 0–100. The overall TEA score is the simple average of all three, with no weighting between axes.
Incumbents (sitting members of Congress) are scored on their voting record, public actions, and measurable outcomes. Challengers (candidates without a voting record) are scored on their published platform, responses to the TEA Questionnaire, and optional pledge commitments.
The scoring system uses a hybrid approach: some criteria are Binary (yes/no, full points or zero) while others are Scaled (proportional, based on the degree of performance). Scores have a zero floor, meaning no criterion can produce negative points.
Scope: This rubric covers federal candidates (U.S. Congress) for the 2026 cycle. It is designed to scale to state and local races in future cycles.
Incumbent Rubric
Members of Congress are scored on their record. Each axis contains six criteria totaling 100 points.
| # | Criterion | Type | Pts | Formula | Data Source |
|---|---|---|---|---|---|
| T1 | Constituent access events | Scaled | 25 | min(25, open_events × 5). Must be open to all constituents (not invitation-only or donor events). 1 event = 5pts, 5+ = 25. | Town Hall Project, official schedule, local press, TEA PAC tracking |
| T2 | Financial disclosure compliance | Binary | 15 | Filed all required disclosures on time and without significant late amendments. Yes = 15, No = 0. | Senate/House disclosure databases, STOCK Act filings |
| T3 | Individual stock trading | Binary | 15 | Does not trade individual stocks while in office. Index funds and blind trusts are fine. Clean = 15, Traded = 0. Red flag if SEC referral or insider trading investigation exists. | Financial disclosures, Capitol Trades, Quiver Quantitative |
| T4 | Public vote explanations | Scaled | 15 | Three sub-criteria, 5pts each: (a) press releases or statements on major votes, (b) regular constituent newsletter with vote explanations, (c) voting rationale page on official website. | Official website, newsletters, press releases |
| T5 | Campaign finance transparency | Scaled | 15 | Three sub-criteria, 5pts each: (a) no super PAC coordination or dark money support, (b) donor disclosure beyond legal minimum, (c) small-dollar donations exceed 30% of total raised. | FEC filings, OpenSecrets |
| T6 | Constituent responsiveness | Scaled | 15 | Three sub-criteria, 5pts each: (a) published response time commitment for inquiries, (b) publicly available constituent services data or annual report, (c) no documented pattern of unresponsiveness or record withholding. | Office reports, news coverage, TEA PAC inquiry audit |
| # | Criterion | Type | Pts | Formula | Data Source |
|---|---|---|---|---|---|
| E1 | Legislative follow-through | Scaled | 20 | Of bills sponsored, what percentage advanced past introduction (committee hearing, markup, floor vote)? (advanced / sponsored) × 20. Minimum 2 sponsored bills required; fewer = pro-rated. | Congress.gov, GovTrack |
| E2 | Vote attendance | Scaled | 15 | (votes_cast / total_votes) × 15. Documented medical/family absences do not count against. | Congress.gov vote records |
| E3 | Outcome-oriented legislation | Scaled | 20 | Of bills sponsored, what percentage include sunset clauses, measurable performance targets, mandatory program evaluation, or reporting requirements? (qualifying / total_sponsored) × 20. | Bill text analysis (congress.gov) |
| E4 | Waste identification & oversight | Scaled | 20 | Two sub-criteria, 10pts each: (a) votes supporting GAO/Inspector General audit recommendations, (b) co-sponsorship of bills that consolidate duplicative programs or require performance review. Scored proportionally against opportunities available. | GAO reports, IG reports, bill co-sponsorship, voting records |
| E5 | Deficit impact | Scaled | 15 | Net CBO-scored impact of legislation the member voted for. Net deficit-reducing = 15, deficit-neutral = 10, net deficit-increasing = 0–7 (scaled inversely by magnitude). | CBO score reports, voting records |
| E6 | Earmark discipline | Scaled | 10 | If earmarks requested: documented, publicly justified district needs = 10, partially justified = 5, no public justification = 0. If no earmarks requested: 7 (neutral). | Congressional earmark disclosures, public justification filings |
Affordability is split into two components: 70% actions/votes (what the member did) and 30% outcomes (what actually happened in their district). The 70/30 split reflects that outcomes are hard to attribute to a single legislator, while still holding members accountable for real-world results.
| # | Criterion | Type | Pts | Formula | Data Source |
|---|---|---|---|---|---|
| A1 | Tax burden on working families | Scaled | 15 | Net effect of tax votes on households earning < $150K. Consistently tax-reducing = 15, mixed = 8, consistently tax-increasing = 0. | CBO distributional analysis, Tax Policy Center, voting records |
| A2 | Housing supply & affordability | Scaled | 15 | Votes and sponsorships for bills increasing housing supply, reducing zoning barriers, funding affordable housing construction. Scored 0–15 based on count and significance. | HUD-related legislation, voting records, bill tracking |
| A3 | Healthcare cost action | Scaled | 15 | Four sub-criteria, ~4pts each: (a) drug price negotiation, (b) price transparency mandates, (c) coverage expansion or public option, (d) PBM reform or healthcare monopoly breakup. | Voting records, bill sponsorship |
| A4 | Consumer cost-of-living action | Scaled | 15 | Votes/sponsorship on: grocery cost relief, childcare affordability, utility rate protection, energy cost reduction. Penalizes votes for tariffs/trade policies that CBO projects will raise consumer prices. Scored 0–15. Omnibus bills scored with context notes. | Voting records, CBO consumer impact projections |
| A5 | Workforce & economic mobility | Scaled | 10 | Votes/sponsorship for: job training tied to employers, CTE expansion, AI worker protections, apprenticeship programs. Scored 0–10. | Voting records, DOL data, bill sponsorship |
| # | Criterion | Type | Pts | Formula | Data Source |
|---|---|---|---|---|---|
| A6 | District cost-of-living trend | Scaled | 15 | Change in cost-of-living index for the member’s district/state relative to national average during their term. Improved = 15, no change = 8, worsened = 0–5. | BLS regional price data, Census ACS, HUD Fair Market Rents |
| A7 | Constituent economic health | Scaled | 15 | Change in median household income relative to cost of living in district during term. Improved = 15, flat = 8, declined = 0–5. | Census ACS, BLS wage data, regional CPI |
Note on outcomes: Outcomes are hard to attribute to a single legislator. The 30% weight reflects that reality. These metrics capture whether a member’s district is actually getting more affordable, while the 70% action weight ensures credit goes to effort.
Red Flags
Red flags are displayed as a visible badge on the candidate card, separate from the numerical score. They do not reduce the score below zero but signal issues voters should know about.
| Flag | Trigger | Display |
|---|---|---|
| Ethics violation | Formal ethics complaint sustained, censure, or reprimand | Red Badge |
| Insider trading referral | SEC referral or DOJ investigation for trading on nonpublic info | Red Badge |
| Campaign finance violation | FEC penalty or fine for campaign finance law violation | Red Badge |
| Felony indictment | Indicted on felony charges while in office | Red Badge |
Challenger Rubric
Challengers have no voting record. They are scored on their published platform and responses to the TEA Questionnaire. The base score runs 0–90 per axis, with an optional pledge bonus of +10 per axis (up to 100).
Challenger cards are clearly labeled: “Scored on platform & questionnaire” to distinguish them from incumbents scored on voting record. Non-respondents display: “Questionnaire not returned. Score: 0.”
Questionnaire Structure
The questionnaire maps directly to the incumbent criteria. For each axis, 6 questions correspond to the 6 incumbent criteria. Each question is worth up to 15 points (6 × 15 = 90 per axis).
Scoring Scale Per Question
| Response Quality | % of Points |
|---|---|
| Specific, measurable commitment with mechanism | 100% |
| Directionally aligned but vague on mechanism | 60% |
| Acknowledges issue, no commitment | 25% |
| No response or opposes TEA position | 0% |
Sample Questions (Transparency Axis)
- 1. Matches T1
“How many open, publicly accessible town halls or constituent forums will you commit to holding per year? Describe the format.”
- 2. Matches T2
“Will you file financial disclosures ahead of legal deadlines and make them available on your website? Describe your plan.”
- 3. Matches T3
“Will you divest from individual stocks and use only index funds or blind trusts while in office?”
- 4. Matches T4
“How will you communicate your voting rationale to constituents? Describe specific channels and frequency.”
- 5. Matches T5
“Describe your campaign finance approach. Do you accept dark money PAC support? What percentage of your fundraising comes from small-dollar donors?”
- 6. Matches T6
“What is your plan for constituent services responsiveness? Will you publish response time data?”
TEA Pledge Bonus (+10 Per Axis)
Candidates can voluntarily sign the TEA Pledge to earn 10 bonus points per axis (pushing the maximum from 90 to 100). Not signing is not penalized.
“I will hold at least 4 open, publicly accessible town halls per year and publish all financial disclosures within 30 days.”
“I will include sunset provisions or measurable performance criteria in every bill I sponsor and publish an annual legislative outcomes report.”
“I will prioritize legislation that measurably reduces the cost of housing, healthcare, and essentials for working families, and I will not vote for trade policies that CBO projects will raise consumer prices without documented offsetting benefit.”
Endorsement & Score Display
Endorsement threshold: A candidate needs a 60% overall TEA score to earn “TEA Endorsed” status. Endorsement is automatic based on score, not a separate editorial decision. If no candidate in a race hits 60%, the race is labeled: “No TEA endorsement in this race.”
Score Display Labels
| Range | Label | Cup Behavior |
|---|---|---|
| 90–100 | TEA Certified | Full cup, heavy steam |
| 70–89 | Strong | Three-quarters, light steam |
| 50–69 | Mixed | Half full, no steam |
| 25–49 | Weak | Quarter full, no steam |
| 0–24 | Failing | Tiny splash, no steam |
Overall Formula
Overall TEA Score = (T + E + A) / 3
Simple average. No weighting between axes. Each axis contributes equally.
Update Cycle
Non-election years: Quarterly updates.
Election years: Monthly updates (January through November).
Every update includes a published changelog documenting what changed and why. Future plans include per-candidate detail pages with raw data points for full transparency.
Data Sources
TEA scores rely on publicly available data. The primary sources include:
| Category | Sources |
|---|---|
| Voting records | Congress.gov, GovTrack |
| Financial disclosures | Senate/House disclosure databases, Capitol Trades, Quiver Quantitative, STOCK Act filings |
| Campaign finance | FEC filings, OpenSecrets |
| Fiscal impact | CBO score reports, Tax Policy Center |
| Government oversight | GAO reports, Inspector General reports |
| Constituent access | Town Hall Project, official schedules, local press, TEA PAC tracking |
| Economic outcomes | BLS regional price data, Census ACS, HUD Fair Market Rents, regional CPI |
| Bill text analysis | Congress.gov full text, committee reports |
Frequently Asked Questions
A grading system that demands transparency from politicians must itself be transparent. Publishing every criterion, formula, and data source means anyone can audit our grades, challenge our conclusions, and hold us to the same standard we hold candidates to.
We believe transparency, efficiency, and affordability are equally important pillars of good governance. Weighting one over another would inject editorial bias. The equal weight keeps the system simple and defensible.
Omnibus bills contain mixed provisions. When a member votes on an omnibus bill, we score the vote and attach a context note explaining the relevant provisions. A “yes” vote on a bill with both cost-saving and cost-increasing measures is noted as such, not treated as purely positive or negative.
No. The floor is zero on every criterion. Egregious behavior triggers a red flag badge (visible on the candidate card) rather than a negative score. This keeps the math clean while still flagging serious issues.
They are scored on separate rubrics. Incumbent scores are based on actual record. Challenger scores are based on stated commitments and platform. Each candidate’s card clearly indicates which rubric was used. We do not equate promises with actions. Voters can see the difference in the card labels.
Good. That means the transparency is working. Every data point and criterion is published here. If you believe a score is wrong, contact us with specific evidence and we will review it. We publish a changelog with every update.
Yes, but it is bounded. Most scoring is algorithmic (data in, score out). Human editorial review applies only to edge cases, such as interpreting mixed omnibus votes or evaluating the quality of a challenger’s questionnaire response. Every editorial judgment is documented.
Ready to see how candidates score? Check out the TEA Grades.
View TEA Grades