Political Stance Arena
Methodology
How arena datasets are built, what they can support, and what they cannot. Published so any number shown in the arena can be traced back to a stated rule.
01 · What this data is
World Stance verified-user data. This is public platform preference among self-selected, identity-verified users — not a scientific poll and not an official election result. Methodology
Arena results describe the preference of World Stance users who chose to take a stance on a given race. Participation is self-selected: users decide which races to open and whether to answer. That makes the data a measure of platform sentiment, not a probability sample of an electorate, and it is not weighted to demographic targets.
02 · The two datasets
Every race maintains two independent tallies, both always visible:
- Local / In-State — stances from users whose verified location matches the race geography.
- National — stances from any identity-verified U.S. user, regardless of location.
A stance always counts nationally. It counts locally only when the verified location matches. An Arizona user taking a stance on the Arizona Governor race appears in both tallies; a California user on the same race appears in National only.
Percentages are always derived from the same counts that produce the displayed sample size, so a published share and its N can never disagree. Shares include users who selected Neither / Undecided in the denominator.
03 · Identity and location verification
Only signed-in accounts can take a stance in the arena. Anonymous stances are never accepted, and the dataset a stance belongs to is decided on the server from the account's stored location — a client cannot assert which state it is answering from.
Current location method: self-declared, stored on the profile. A user selects their state once, either in their profile or through a prompt shown at their first stance. It is not inferred from IP address or network signals, so a stance cannot be relocated by changing device or network — but it is also not documentary proof of residence, and this page will be updated if and when a stronger method replaces it.
The state is recorded onto the stance itself at the moment it is taken, rather than read back from the profile later. A user who moves and updates their profile does not retroactively change any tally they have already contributed to.
Anti-fraud measures include one stance per account per race, enforced by a database constraint rather than by the interface, and stances that cannot be attributed to an account are rejected outright. A recorded stance is locked and cannot be edited or deleted by the user, so a published tally cannot be quietly revised. Stances removed by later integrity review are removed from both datasets, which is why a tally can move down as well as up.
04 · Display rules
A candidate's percentage takes the emphasized treatment only when it is both at or above 50% of all stances and ahead by at least 5 points. Below either threshold both candidates are shown in the same neutral treatment, so a close race never reads as settled.
Local datasets below 10,000 stances carry a small-sample caution. Candidate order within a race is fixed and never sorted by standing.
05 · Equal treatment for campaigns
Both sides of a race render from one template with identical limits. Surplus assets are truncated by the same rule on both sides, so no campaign can occupy more column space than its opponent:
- 1 hero video
- Up to 4 additional videos
- Up to 5 “Why Support” cards, 220 characters each
- Up to 3 issue / response modules
- Up to 3 related stance question links
Campaigns control only their own side. Ranking and visibility are never driven by party. Every creative unit carries a persistent “Paid for by” disclosure naming the authorized committee.
06 · Geography definitions
Curated presets are editorial groupings maintained by World Stance and published here in full. Membership is never inferred from results:
Battlegrounds
States World Stance tracks as competitive this cycle.
AZ · GA · MI · NV · NC · PA · WI
Sun Belt
Southern and southwestern growth states.
AZ · CA · FL · GA · NM · NV · NC · SC · TX
Midwest
Great Lakes and upper-plains states.
IA · IL · IN · KS · MI · MN · MO · NE · ND · OH · SD · WI
Northeast
New England and mid-Atlantic states.
CT · DE · MA · MD · ME · NH · NJ · NY · PA · RI · VT
County and media-market filtering is published only where verified density supports it, and is not yet enabled.