Methodology
Our alignment engine is deliberately simple, so it can be inspected and challenged.
1. Questions
Every question is a plain-English statement about a real policy area — climate, housing, tax, health, education, immigration, civil liberties, foreign affairs and more. Users answer on a five-point scale from strongly disagree to strongly agree.
2. Party positions
Each question corresponds to a public policy position. Rather than tagging a party with a single label, we store their stance as a probability distribution across the five answer options — the five numbers always sum to 1.0. This lets us capture certainty as well as direction.
2a. How the party spread is decided
A value of 1.0 on one option means the party has an unambiguous, clearly stated public position. A split like 0.8 / 0.2 means a strong lean with a mild hedge — the platform points one way, but there is nuance, internal dissent, or softer language in official materials. A 0.6 / 0.4 split means a moderate lean where the direction is discernible but genuinely contested or conditional. Weight on "neutral" means the party has no clear public position or deliberately straddles the issue. Distributions are seeded from official party platforms, media releases and voting record, then reviewed for balance. When the AI policy engine ingests a new verified source it proposes updates to these five numbers rather than flipping a single label, and every change is logged with a source and date on the updates page.
3. Scoring
For every answered question we compute the distance between your answer and each party's position, then convert distance to a similarity score between 0% and 100%. Your overall alignment with a party is the average across all answered questions.
4. Confidence
Confidence rises with the number of questions answered. QuickDNA (the free 20-question assessment) gives an indicative snapshot. The Complete DemocracyDNA continues the same assessment in more depth for a higher-confidence result.
5. AI's role
AI never creates opinions. It monitors official sources for policy changes, drafts summaries, and updates the policy database. Every change is logged with a source and confidence score on our public updates page. Low-confidence changes require human review before they go live.
6. What this is not
This tool does not tell you who to vote for. It only shows where your answers sit relative to published party positions. Your vote is yours.