Method
How the evidence was built
The case study treats Formula 1's 2026 rules as a stakeholder communications problem, not a technical one. Everything here can be checked against the evidence page.
Scope
- Window: February 1 – September 30, 2026. Events after September 30 are context only.
- Question: every statement is coded against one proposition: The 2026 regulations have been bad for Formula 1. A speaker who agrees is critical; one who disagrees defends the rules.
- Unit: one signal per stakeholder per source item. In press conferences, everything a person said about the rules in that session is one signal, so signal counts measure people speaking, not answer length.
Sources
- Backbone (124 signals): all 65 official FIA press conference transcripts published for the window, Australia (March 5) to Azerbaijan (September 26). Every participant is logged, including those who said nothing about the rules, so silence is visible but never treated as a stance. No transcripts were published for February testing.
- Supplementary (56 signals): statements and press releases from the FIA and Formula 1, and interviews in established outlets. These fill the gaps the press conferences leave: the institutions, the manufacturers and pre-season testing. Every URL was fetched and checked; verbatim excerpts are under 15 words.
- Fans: RaceFans "Rate the race" reader polls (15 Grands Prix in 2026, plus the 2025 season as a baseline), and 2,861 public comments from 84 RaceFans threads, sampled up to 60 per thread. Reddit was not used: it blocks automated access and requires prior approval for its data interface. Other forums block automated access. Only aggregates are published, with no comment text, usernames or links to individual comments.
Frames
Up to three frames per signal, coded for the argument the speaker advances, not one they rebut. Seven come from a general public-affairs taxonomy; five are specific to motorsport.
| Frame | Origin | Definition | Example |
|---|---|---|---|
| Spectacle and entertainment | Motorsport extension | The racing is judged as a show for fans and broadcasters: overtaking, wheel-to-wheel battles, excitement, whether races are worth watching. | A driver says fans at home cannot understand why cars slow down on the straights, so the racing looks artificial. |
| Driver skill and sporting purity | Motorsport extension | Whether the cars reward driver ability, flat-out driving, and racecraft, or reward energy management and software. | A driver says qualifying is no longer about attacking every corner because they must lift to recharge the battery. |
| Road relevance and sustainability | Motorsport extension | The link between the rules and road-car technology, electrification, sustainable fuels, or the sport's environmental commitments. | A manufacturer says the 50-50 electrical split is why it entered the sport, because it mirrors its road-car program. |
| Competitive balance | Motorsport extension | Whether the rules favor particular manufacturers or teams, including engine performance gaps, loopholes, and catch-up mechanisms. | A team principal argues that a rival engine maker found an interpretation of the compression-ratio rule that others cannot match. |
| Stability and patience | Motorsport extension | Arguments against reacting quickly to criticism: give the rules time, teams will converge, new rules always take time to settle, knee-jerk changes are costly. | The commercial rights holder says every new rule set draws early criticism and the field will close up. |
| Safety | General | Physical safety of drivers, marshals, or spectators: closing speeds, unpredictable deceleration, crash risk. | A driver warns that large speed differences between a car harvesting and a car deploying could cause a serious accident. |
| Fairness | General | Fair treatment and due process in how rules are made, applied, or enforced: consistency of stewarding, whether all parties had a say. | A team principal says the rule adjustments were agreed without the smaller teams being properly consulted. |
| Feasibility and implementation | General | Whether the rules or proposed changes can practically work: technical readiness, reliability, cost-cap limits, lead times for changing hardware. | An engineer says changing the energy deployment limits mid-season is possible in software, but hardware changes need a year. |
| Innovation and competitiveness | General | The rules as technical ambition: new technology, engineering challenge, attracting manufacturers, keeping Formula 1 at the leading edge of motorsport. | The FIA says the regulations are the most advanced power units ever raced and have brought new manufacturers into the sport. |
| Institutional trust and accountability | General | Whether the governing body and commercial rights holder are transparent, listen to criticism, and are accountable for decisions, including how they communicate. | A driver says the sport should listen to drivers rather than edit their comments out of official videos. |
| Business and economic climate | General | Commercial effects: audiences, broadcast value, sponsors, manufacturer investment, team finances, the value of the sport as a business. | An executive points to record attendance and viewing figures as evidence the product is healthy. |
Stance and confidence
| Stance | Definition |
|---|---|
| Critical of the rules | The signal says or clearly implies the 2026 rules have made the sport worse. |
| Defends the rules | The signal says or clearly implies the 2026 rules are good, working, or improving. |
| Mixed or conditional | The speaker sends both signals in the same statement, or judges the rules acceptable only if specific changes are made. |
| Noncommittal | The speaker addresses the rules without judging them, or deflects. |
| Procedural only | The signal concerns process or scheduling with no evaluative content. |
A stakeholder's overall stance is a weighted average of their signals that take a position (critical +1, mixed 0, defends −1). Higher-commitment contexts and more recent signals weigh more. Confidence follows a calibration table:
- High: 3+ consistent signals, at least 1 explicit, recent, no contradicting signals.
- Moderate: 2+ signals with partial consistency, or one strong explicit signal with indirect confirmation; minor contradictions allowed.
- Low: 1–2 signals, indirect evidence, or partially conflicting signals.
- Insufficient: 0–1 weak signals; any stance would be speculative. Silence alone is never a stance.
Noncommittal signals stay in the log but are left out of averages and volume counts. Headline results were checked with and without them.
Reliability
All coding was done by an automated coder, Claude, following written protocols. A second, independent automated coder from a different model family, briefed separately and blind to the first coding, recoded random samples. This measures agreement between two automated coders. It is not human validation.
| Measure | Stakeholder signals (55-item sample) | Fan comments (286-item sample) |
|---|---|---|
| Whether a remark addresses the rules (Cohen's kappa) | 0.56 | 0.71 |
| Stance agreement | 86% (kappa 0.80) | 80% (kappa 0.69) |
| Critical vs everything else (kappa) | – | 0.76 |
| Frame overlap (mean Jaccard) | 0.75 | 0.74 |
Two problems surfaced and were fixed before analysis. First, the first-pass coder was more inclusive than the second about what counts as a remark on the rules. In response, averages and volume counts use only signals that take a position. Second, two of six fan-coding batches applied a rule differently. Four harmonisation rules were written and 514 comments in the affected threads were recoded by a single coder. Fan stance kappa was 0.67 before recoding and 0.69 after.
Derived measures
- Movement: average stance by group across four periods (testing and opening rounds; the refinements; summer; run-in) and by race round.
- Frame divergence: Jensen-Shannon divergence between groups' frame mixes, with bootstrap intervals and a random-split baseline.
- Response timing: weekly critical remarks set against official actions and statements, read as an event study.
- Race ratings: pooled vote averages for 2026 against 2025, with a bootstrap interval.
Limitations
- Press conferences capture what people said when asked. Panel rotation means a driver can be silent for weeks.
- The FIA and Formula One Management rarely speak in that format, so their stance rests on fewer, supplementary signals. Formula One Management's has 6.
- Honda and General Motors have no evaluative public record in the window.
- Fan data comes from one English-language site with a self-selected, engaged readership.
- The reported removal of a Verstappen remark from Formula 1's social media has a single source (Autosport) and is presented as reported.
- The stance-by-influence view in the original plan is not shown. Its influence scores are analyst judgment still under review.
