Threshold
Research
Services
A public instrument for measuring where the Republican coalition holds, where it breaks, and where it tips.
First cycle · October 2026 – February 2027
G2 Analytics · Confidential draft
At a glance
The party is at a threshold. Nobody is measuring it. We will.
What
A standing, public instrument that tests the coalition’s commitments under pressure and publishes the result on a schedule, with its method.
Who
Seven layers of the Republican coalition, five edge audiences and a Democratic mirror panel: 2,000-plus panelists, re-contacted after the vote.
When
Baseline fielded the week before November 3. Post-election read by November 20. State of the Threshold published February 2, 2027.
Turbulent zone
Split or unstable. Declared open.
Threshold
Holds, then gives way.
Laminar zone
Holds everywhere. The doctrine.
Every commitment the coalition is assumed to hold lands in one of three zones. The thresholds are where elections turn.
Not a poll
A poll measures opinion at rest. Threshold measures what a commitment survives under attack, in a trade-off, and from a hostile messenger.
Not a platform
The doctrine is found, not decreed. It updates when the data does, in public, on a schedule.
Not for sale
No funder sets the questions or sees results before the outside reviewer. The report ships on its date whatever it says.
01 — The offer
Every faction has a theory. None of them has a measurement.
Threshold is a standing, public instrument that tests the Republican coalition’s commitments under pressure, publishes the result on a schedule, and shows its work. It replaces the oldest defect in politics, the inference a few people make about what the coalition wants, with a measurement anyone can inspect.
It is not a poll. A poll measures opinion at rest and gets an answer that is cheap to give and cheap to abandon. Threshold measures what happens to a commitment when it costs something: when it is attacked by the other side’s best argument, traded against a second commitment the same voter holds, and delivered by a messenger the voter dislikes. Reactions are captured moment to moment on a live panel and modeled across the full electorate.
The first act is to measure the 2026 electorate before it votes and again after. The midterm is the last moment the current coalition exists in something like its original form, and nobody is recording it. That baseline is the founding asset of the project and the only honest before-and-after account of what happened.
The promise
We will stop telling voters what they believe, and start finding out.
2,000+
panelists across every layer of the coalition, re-contacted after the vote
4
pressures on every commitment: at rest, attack, trade-off, messenger
Feb 2
2027: the first State of the Threshold report, published with its method
02 — The suite
Six services. One instrument. One map.
Everything below runs on the same panel, the same pressure battery and the same zone rules, so every client reads from the same map. Pricing is scoped per engagement.
Flagship · public · annual
State of the Threshold
The published map: laminar zone, turbulent zone, every threshold with its break point, the method, the validation and year-over-year movement. Founding funders get first read and credit.
Feb 2, 2027, then annually
Founding dataset
The Baseline
Pre-election and post-election reads on the same 2,000-plus panelists across every layer. Topline hold scores by layer within a week of each close.
Fielded Oct 26 – Nov 20, 2026
Campaigns · 2027–28 field
Candidate Read
A candidate’s platform measured against the map: which positions sit in the laminar zone, which are thresholds, and exactly where their own layers give way. The same method for everyone.
From Q1 2027 · 3 weeks
Campaigns · committees · advocacy
Pressure Test
One message, ad or commitment through the four pressures on the standing panel, with moment-to-moment reaction curves, break points by layer and messenger effects. The fastest read in the suite.
Rolling · 10 days
Subscription · donors · press
Layer Monitor
Quarterly re-contact of the panel on a fixed core of commitments. Shows which layers are separating and which thresholds are moving between annual reports.
Quarterly from Q2 2027
Donors · party leadership
Elite Gap Study
Twenty to thirty structured interviews with donors, chairs, consultants and members, asked to predict the map before they see it. The gap between prediction and measurement is the headline exhibit.
Alongside each Baseline
03 — The flagship report
State of the Threshold, February 2027.
The first report is the founding document of the project: not a manifesto, but a map. It is published with its method, repeated every year, and compared against the year before. Its chapters are fixed now so no one can add or remove one after seeing the numbers.
The map
Every tested commitment placed in its zone, with hold scores by layer. One page per commitment.
The laminar zone
The commitments that held everywhere under every pressure. The doctrine as measured, not as decreed.
The thresholds
For each contested commitment: the break point in the attack, the layer that breaks first, and the trade-off that beats it.
The turbulent zone
The questions the coalition is split or unstable on, declared open, with what it would take to find out.
Movement
The same panel before and after November 3: what the vote moved, by layer, and which thresholds shifted more than 10 points.
The elite gap
What donors, chairs and consultants predicted the map would show, set against what it showed.
Method and review
Panel design, the pressure battery, the zone rules as frozen before fielding, the replication check, and the outside reviewer’s signed note or dissent.
Founding funders receive the report two weeks before publication. The outside reviewer receives it before they do.
04 — Audiences
Seven layers of one coalition, each read on its own.
The population is the Republican coalition as it exists in 2026: registered voters who identify as Republican, lean Republican, or voted Republican in 2020 or 2024. Layers are assigned from a short screener rather than self-description, so a voter can sit in more than one, and every stimulus is read within every layer.
| Layer | Defined by | Baseline n |
|---|---|---|
| Business conservative | Taxes and regulation top priority; college-educated; suburban | 300 |
| Evangelical / social conservative | Religion very important; abortion or religious liberty top priority | 300 |
| National security hawk | Defense or foreign policy top priority; supports Ukraine aid or confronting China | 250 |
| Libertarian-leaning | Government too big; socially permissive; low institutional trust | 200 |
| Populist / America First | Immigration and trade top priority; first GOP vote 2016 or later; non-college | 400 |
| Disaffected / low-propensity | Voted 2024 but not 2022; low trust in both parties | 300 |
| Persuadable independent | No party ID; voted Republican at least once since 2016 | 250 |
| Core sample | Oversampled in the three moving layers; weighted back to the coalition’s shape from the voter file | 2,000 |
Re-contact built in
Every panelist is recruited with consent to be re-contacted, so the post-election read runs on the same people and the map carries direction, not just position.
Small cells are fine
The first cycle’s job is to show the map works, not to call a race. Depth by state is a second-cycle goal.
04 — Audiences, continued
Five edges, and the elites who think they already know.
Each edge audience is added for a specific reason and is small enough not to move the budget. The mirror panel is the calibration: a commitment Democrats also hold is a consensus, not a Republican threshold.
| Audience | Why it is in the study | What it answers | n |
|---|---|---|---|
| Trump-to-nothing voters | Voted 2024, unlikely to vote 2026. The energy the coalition is shedding. | What brings them back; what pushes them out | 200 |
| Young Republicans, 18–29 | Least Reagan-era memory; most turbulent reads. | Whether the laminar zone outlives this cycle | 200 |
| Hispanic Republicans and leaners | Fastest-growing layer; most sensitive to the immigration trade-off. | Border vs. legal immigration; messenger effects | 200 |
| Suburban swing women | Left in 2018 and 2022; partly returned in 2024. | Where abortion and leadership commitments break | 200 |
| Democratic mirror panel | Same battery, run on Democrats. | Calibration: consensus vs. threshold | 300 |
Elite audiences
Twenty to thirty interviews with donors, state chairs, consultants and sitting members, asked to predict the map before they see it. Researched qualitatively and reported against the measurement. The gap is the clearest proof that the inference was never checked.
Who the instrument serves
In order: candidates and their strategists, who get the map; donors, who get an honest account of what they are funding; the press, which gets a public, method-first read of the party; and voters, who are told what the party found rather than what it decided.
05 — Method
Four pressures, one dial, and rules written down before anyone sees a number.
1
At rest
The commitment delivered plainly by a neutral narrator. The poll-equivalent and the control.
2
Attack
The strongest case the other side would make, written the way a Democratic ad would be. The dial shows where support starts to give.
3
Trade-off
The commitment set against a second commitment the same voter holds. Which one wins is the threshold.
4
Messenger
The same statement from a messenger the layer trusts and one it does not. A commitment that survives from only one mouth is not laminar.
The commitments
Twelve to fifteen statements the coalition is assumed to hold, drawn from the 2024 platform, the loudest factions’ claims and the places the layers collide. Candidates for the first battery:
- Lower taxes are worth a larger deficit
- Tariffs are worth higher prices
- Secure the border before any legal-immigration reform
- Stand with Ukraine against Russia
- Confront China even at a cost to trade
- Protect Social Security and Medicare as they are
- Abortion should be decided by the states
- Back the party’s leader even when he is wrong
- Cut federal agencies sharply
- Cut aid abroad before any cut at home
Protocol and scoring
Panelists see six commitments each, so every commitment gets roughly 800 reads per pressure across layers. After each block, a short survey: support on a 0–100 scale, willingness to accept a cost, and whether they would stay in the coalition if the party abandoned the commitment.
Scoring uses G2’s validated reaction-curve method: baseline, peak and trough, time-to-break, recovery and the gap between layers. A commitment holds in a layer when 60% or more stay above the dial midpoint through the attack and within 10 points of their at-rest support afterward.
Zone rules, frozen before fielding
Laminar: holds in every layer under attack and the disliked messenger, and wins its trade-off in at least five of seven layers.
Turbulent: holds in fewer than four layers, or layers disagree by 25 points or more, or it loses its trade-off in most layers.
Threshold: everything in between, published with its break point, first-breaking layer and the trade-off that beats it.
Two checks before publication: a split-sample replication on 20% of the battery, and an outside review of the zone rules by a pollster not paid by the project. Both are published with the report.
06 — First cycle
The only hard date is November 3.
Three deliverables, and the first one has to be in the field inside two weeks. The timeline runs October 17, 2026 to February 2, 2027; the dashed line is election day.
Deliverable 1
Pre-election baseline: the full battery on 2,000-plus panelists. Topline hold scores by layer within a week of close.
Deliverable 2
Post-election read: the same battery on the same panelists. Shows what the vote moved.
Deliverable 3
State of the Threshold 2027: the public report, the method, the validation and the movement. Repeated annually.
If funding is not signed by October 24, the baseline shrinks to the three moving layers and six commitments rather than slipping past the election. The second cycle begins with the Layer Monitor in Q2 2027 and the next full battery in autumn 2027.
07 — Governance
An instrument that can be captured is just a more expensive microphone.
These rules are fixed before the first dollar is taken. They are what make the measurement worth more than the opinion of whoever paid for it.
Published on a schedule, not on a verdict
The report date is set at funding. It ships on that date whatever it says.
Methods are public
Panel design, commitment list, the four pressures, zone rules and weighting are published in enough detail to replicate.
Rules frozen before fielding
Hold-score threshold and zone rules are locked in writing before the first panelist is recruited.
No single funder owns it
A founding funder gets first read and credit. No funder sets the commitment list or sees results before the outside reviewer.
Outside review
A named pollster not paid by the project signs the report or publishes a dissent with it.
The data belongs to the instrument
Held by G2 for the first cycle; ownership for later cycles decided before cycle two, not after.
Candidates are measured, not barred
Anyone can run on anything. The instrument reports how their positions sit against the map, the same way for everyone.
The data decides
If a candidate’s bet against the map is right, the map will show it and the doctrine will move. If it is wrong, the map will show that too.
Why G2
G2 Analytics runs moment-to-moment dial testing for political campaigns and committees. The reaction-curve scoring the instrument uses is the same method G2 delivers to clients today, validated against real deliverables rather than built for this project. What is new is the stimulus design, the standing panel and the publication, not the measurement.
08 — How to engage
Three ways in. One is open for seven more days.
Founding funder
Closes Oct 24
Funds the 2026 Baseline. First read of every result, credit in State of the Threshold, a seat at the method review. Closes October 24 so the baseline can field before the election. No thumb on the scale.
Adopting campaign
From Q1 2027
Runs on the laminar zone and is measured against it. Receives the Candidate Read and the Pressure Test at member rates and works from the same map as every other adopter.
Subscriber
From Q2 2027
Donors, committees and press who want the Layer Monitor between annual reports. Quarterly, on the standing panel, with the previous quarter to compare against.
What we need by October 24
Funding source and scope. Full battery, or the three-layer fallback.
The name. It goes on the panel consent form, so it precedes recruitment.
An outside reviewer. One pollster, named in the report, who agrees to the zone rules before fielding.
Panel source. Opt-in panel with voter-file match, or recruitment from a partner list.
Pricing is scoped per engagement. Contact details go here before this leaves the building: [NAME], [EMAIL], [PHONE].
Every previous attempt to remake the party started with a set of beliefs and went looking for voters to hold them.
This starts with the voters.
Threshold · A G2 Analytics instrument
Confidential draft, October 2026. Not for distribution.