Constituent Alignment Score: Methodology
Methodology version 12. Every published score is stamped with the version that produced it.
What this score is — and isn't
The Constituent Alignment Score measures one question: how well is this person doing for their specific constituents?An elected official's loyalty belongs to the people of their own district or state — not to donors elsewhere, and not to like-minded politicians representing other places. The score rewards serving your own voters and penalizes loyalty that points somewhere else.
It is a deterministic function of Tier 1 public records under the published weighting below— arithmetic, not an opinion. We assert the inputs and the formula, never a verdict about a person's character. Every metric links to its own source, and you can reweight the formula yourself on any profile and watch the overall recompute — the underlying facts never change.
Excluded by design
Voting with (or against) like-minded people from other places is neither rewarded nor punished. The following are deliberately not inputs to this score:
- Party-unity / party-line agreement rates
- Ideology scores (DW-NOMINATE and similar)
- Agreement with any bloc, caucus, or coalition
- Endorsements
Only the relationship between the official and their own constituents counts.
The four components and their weights
Four components feed the overall. Every input is a Tier 1 government record already flowing through the data pipeline.
| Subscore | Default weight | What it measures |
|---|---|---|
| In-State Funding | 30 | Who pays for their campaign — their own constituents or outsiders? The share of itemized individual contribution dollars (FEC Schedule A) that comes from donors inside the candidate's own state. PAC and party money is excluded here: it has no home district. Donor-linked influence is the Donor Independence component's job. Scored only when donor-state data covers at least half of the candidate's itemized receipts, so nobody is graded against a thin sample. |
| Roll-Call Participation | 20 | Do they show up to vote at all? The share of recorded roll calls in the current Congress where the member cast a definite yea or nay. Scored only when vote ingestion covers at least 90% of the chamber's roll calls, so nobody is graded against an incomplete denominator. |
| STOCK Act Discipline | 15 | Are they serving themselves on the side? Starts at 100 and deducts for each disclosed trade filed outside the 45-day statutory window and each trade whose timing overlaps legislation naming the issuer. Counts of factual observations from official disclosures — never characterized as wrongdoing. No disclosed trades is a clean 100: nothing to penalize is a fact, not a gap. |
| Donor Independence | 35 | When top out-of-state donors want something, do their votes follow? The inverse of the donation-weighted alignment rate between the member's votes and outcomes favoring their largest non-constituent donors (influence engine, issue 010). High = votes don't track outside money. |
The overall is a weighted average of whichever components have data; weights are renormalized across the present components, so a missing component is dropped rather than counted as zero. With fewer than all 4 present, the score is badged provisional — N of 4 components, and is capped at 99: a perfect 100 is reserved for a perfect record across all 4 measured dimensions.
How each component is computed
- In-State Funding = 100 × in-state itemized individual dollars ÷ all itemized individual dollars with a known donor state (FEC Schedule A). PAC and party money is excluded — it has no home district. Scored once at least 25 itemized receipts carry a donor state.
- Roll-Call Participation = 100 × roll calls where a definite yea/nay was cast ÷ recorded roll calls. Scored only with 20+ roll calls ingested and ingestion covering ≥90% of the chamber's roll calls — nobody is graded against an incomplete denominator.
- STOCK Act Discipline starts at 100: −15 per trade filed outside the 45-day statutory window (this deduction class stops at 40), −10 per legislation-timing-overlap flag (class floor 60). No disclosed trades is a clean 100 — nothing to penalize is a fact, not a gap. These are counts of factual observations from official disclosures, never characterized as wrongdoing.
- Donor Independence= 100 − 100 × the donation-weighted alignment rate between the member's votes and outcomes favoring their top out-of-state donors. Alignment is a temporal correlation (votes cast after the money arrived), not a claim of causation.
When the overall is withheld
The Constituent Alignment Score is fundamentally an in-office accountability measure, so we publish a headline number only when all three of the following hold:
- The person holds office (or has an actual record).A pure challenger with no voting record is never given a headline number; their profile shows the available facts under “No voting record to score yet.”
- Their data import is complete — every government source that could apply to them has been fetched and resolved. A score computed off a half-finished import would be misleading, so we withhold it until the import lands.
- At least 2 of the 4 fact-based subscores are backed by real data. Below this floor there is too little record to summarize.
Below the bar, the subscores are still shown individually with no headline number, under “Not enough record to score yet.” Missing data is never imputed as 0 or 50. The score lights up for each person as their import completes and government connectors and community submissions add data.
Reproducibility
Every score row stores the methodology version and the time it was computed. Changing any number on this page bumps the version and triggers a full recompute of every candidate, so no stale-methodology scores linger.
See also our source-tier rules — every metric is a single public-record fact presented under those rules.