Per topic
Individual opinion weight
Every opinion begins with zero weight. For a particular topic, the application adds the importance values of the published fact questions the participant currently answers correctly, then divides that by the importance available across the complete published fact bank.
If a participant answers the same fact more than once, the newest response to that fact replaces the earlier one. Responses to every other fact continue to count. The displayed unweighted score remains available as a check on how much the editorial importance ratings affect the result. Knowledge demonstrated on one topic never carries into another.
Editorial assessment
Fact importance
Each fact has one of three deliberately narrow values: Supporting (1), Significant (2), or Foundational (3). Importance concerns how necessary the fact is for an informed judgement about the proposition—not whether it supports or counters that proposition.
Every fact also records its likely evidential direction: supporting the proposition, countering it, or contextual/mixed. Direction affects presentation order only. It does not alter the correct answer or numerical weight.
Public result
Collective informed opinion
The five ordered response positions are assigned values +1, +0.5, 0, −0.5 and −1 from strongest agreement to strongest disagreement. Each value is multiplied by that participant's topic weight.
The result lies between −1 (No) and +1 (Yes), with zero as Neutral. Zero-weight opinions are counted as registered opinions but cannot move the indicator. A weighted neutral response adds to the denominator without adding a signed value and therefore draws the result toward neutral.
Presentation
Question sequence and review
Unseen facts are shown before review questions. The first eight positions favour editor-designated gateway facts that establish the principal considerations without requiring specialist preparation. The sequence also normally begins with evidence closer to the participant's registered view and introduces challenging evidence progressively. These preferences never remove contrary evidence or change scoring.
After every fact has been seen, review prioritises the least-repeated material so a small group of questions cannot recur indefinitely. Answer choices are shuffled to reduce answer-position cues.
Editorial standards
Quality and correction
Published facts should be relevant to the opinion, have one clearly supported answer, use plausible alternatives and link to an accessible source. A bank should cover the strongest material evidence on every side without manufacturing an artificial balance.
Participants may report material that appears inaccurate, irrelevant, unclear, biased or outdated. A corrected question may be rescored when its answer key was wrong; a retired question remains in private answer history but no longer contributes to current scores.
Question wording, sources, importance and direction remain editorial judgements. The project publishes those judgements and its formula so they can be scrutinised rather than treating them as neutral machinery.
Interpretation
What the result does not prove
Participants are self-selecting, so this is not a representative opinion poll. A high weight demonstrates performance on the selected fact bank; it does not prove expertise, sincerity, sound judgement or that an opinion is correct.
The aggregate describes participating opinion under the published question bank and calculation at that time. Participant counts, source quality, question selection and editorial decisions should all inform how cautiously it is read.