Delegation Cohesion
Some state delegations vote as a bloc. Others barely agree with themselves.
By Dallas AI Metrics, LLC · Analysis as of 2026-08-04 · Computed from 440 roll calls across 526 members · Methodology
Members of Congress are elected by states and districts, but they vote in parties. Most of the time those two loyalties point the same direction. Where they diverge, a state delegation stops behaving like a bloc.
Delegation cohesion measures how tightly the members of one party from one state vote together with their national party. A delegation whose members all hold the party line has high cohesion. One containing both a reliable partisan and a frequent defector has low cohesion, even if every member is nominally on the same team.
What this is and is not measuring
The measure is built from the same party-unity calculation used in the defection index, averaged across the members of each state-and-party group. High cohesion means the group's members individually voted with the national party; low cohesion means at least one member frequently did not.
That is a specific and slightly counterintuitive definition, so it is worth stating what it does not measure: it does not measure whether a delegation votes together as a state. A delegation whose members all broke from their party in the same direction on the same regional issue would score as low-cohesion here, despite being perfectly unified in local terms. This measures alignment with the national party, using the delegation as the unit of observation.
Why delegations differ
Several forces push in different directions, and the ranking below reflects their sum rather than any one of them.
Seat competitiveness. Members from safe seats face little electoral cost for voting the party line. Members from seats their party barely holds have strong incentives to establish independence on visible questions. A delegation containing marginal seats will tend to show lower cohesion for reasons that have nothing to do with regional politics.
Delegation size. Small groups are volatile. In a two-member delegation, one member's idiosyncrasies move the group average enormously; in a large delegation, individual variation averages out. Small delegations will over-populate both ends of the ranking, and that is an artefact of sample size rather than a finding about those states.
Regional interests that cut across party. States with a concentrated dominant industry, a distinctive geography, or a specific federal relationship produce recurring votes where the state interest and the party position genuinely conflict. These are the cases where low cohesion is substantively interesting rather than statistical noise.
| # | State | Party | Members | Cohesion |
|---|---|---|---|---|
| 1 | Louisiana | D | 2 | 98.6% |
| 2 | Missouri | D | 2 | 98.6% |
| 3 | Minnesota | R | 4 | 98.5% |
| 4 | Rhode Island | D | 4 | 98.5% |
| 5 | Washington | R | 2 | 98.2% |
| 6 | Michigan | R | 7 | 97.5% |
| 7 | Minnesota | D | 6 | 97.3% |
| 8 | Delaware | D | 3 | 97.3% |
| 9 | Mississippi | R | 5 | 97.1% |
| 10 | Arkansas | R | 6 | 97.0% |
| 11 | Maryland | D | 9 | 97.0% |
| 12 | Indiana | R | 9 | 96.8% |
| 13 | Hawaii | D | 4 | 96.6% |
| 14 | Nebraska | R | 5 | 96.5% |
| 15 | Illinois | D | 16 | 96.4% |
The low end
Delegations at the bottom of the ranking are worth examining case by case rather than as a group. Low cohesion can mean a genuine regional split from the national party, or it can mean the delegation contains one unusually independent member and several ordinary ones.
Those two situations look identical in a group average and are completely different politically. The individual member records behind each delegation are the only way to tell them apart, and the summary figure should be treated as a pointer toward that question rather than an answer to it.
| State | Party | Members | Cohesion |
|---|---|---|---|
| Maine | D | 2 | 85.5% |
| Alaska | R | 3 | 85.6% |
| Kentucky | R | 7 | 87.1% |
| Wyoming | R | 3 | 90.4% |
| Arizona | R | 6 | 90.9% |
| South Carolina | R | 7 | 91.6% |
| New Jersey | R | 3 | 92.1% |
| Missouri | R | 8 | 92.1% |
| North Carolina | D | 4 | 92.1% |
| Nevada | D | 5 | 92.5% |
Exclusions
A delegation with only one member of a given party is excluded entirely. Cohesion within a group of one is undefined — the single member is, trivially, in perfect agreement with themselves — and reporting such a case as either total cohesion or none at all would be meaningless. A substantial number of state-and-party groups fall into this category and simply do not appear in the tables above.
This exclusion is not a minor technicality. It removes every state where one party holds a single seat, which disproportionately affects small states and minority-party representation in one-sided states. The delegations shown here are the ones large enough to measure, and that set is not a random sample of the country.
Both chambers are pooled, so a state's senators and representatives of the same party are treated as one group. The full rules are in the methodology.