The working behind the quiz.

Every political quiz makes choices. This one makes more than most. Here's what we did, why, and the sources we drew from, so you can argue with us properly.

1 · What this quiz is, and isn't

This is an editorial tool, not a psychometric instrument. It's designed to help you think about your politics in more than the usual two dimensions, and to offer a result with enough character to be worth discussing. It draws on real academic work: Eysenck's two-axis model, Bourdieu's cultural capital, Haidt's moral foundations, the broader Political Compass tradition, but it isn't peer-reviewed, isn't validated against a control sample, and doesn't claim to predict your behaviour.

What we'd say it is good for: surfacing internal contradictions you might not have noticed, giving you better vocabulary for your own positions, and possibly an argument at the pub. What we wouldn't claim: that it diagnoses you, that the archetypes are scientifically real, or that the party percentages tell you who to vote for.

2 · The nine axes

Most political quizzes use two axes: economic left-right and social authoritarian-libertarian. That's the model Hans Eysenck proposed in 1954 and which The Political Compass popularised. It's good but limited: it can't tell apart a free-market progressive from a free-market traditionalist, or a localist socialist from a globalist one.

We expanded to nine. The first two are the classical pair. The other seven are additions designed to capture the parts of your politics that the standard compass misses. The ninth, the planet, was added in October 2026 after the first weeks of results: the Eco Warrior card was going to the most progressive and international respondents whether or not they cared about ecology, because nothing in the quiz measured it.

The economic and social axes are weighted 1.2× the others, because they remain the most consequential dimensions of political identity, even with seven others alongside.

3 · The fourteen archetypes

The archetypes are editorial constructs, not validated personality types. We built them by looking at how the nine axes cluster in recognisable political traditions, both contemporary (the Cosmopolitan Liberal, the Sovereignist) and historical (the Old Socialist, the Templar). Each archetype is defined by a centroid in nine-dimensional space; matching is done by weighted Euclidean distance.

Names like "Templar" or "Bohemian" don't appear in any political science paper. They're labels we chose to feel evocative rather than clinical. A standard k-means cluster analysis on enough real responses would probably produce different, less colourful groupings. We made an editorial choice: better stories beat better statistics for a quiz that's trying to be read, not cited.

The left-of-centre half of the spectrum is more densely populated than the right: we have seven archetypes broadly there (Eco Warrior, Old Socialist, Cosmopolitan Liberal, Anarchist, Dissenter, Steward, Builder) versus three on the right (Templar, Free Marketeer, Sovereignist), plus four shapes that don't fit neatly on either side. This isn't bias. It's a reflection of where Bulletin's network sits, where we'd want the quiz to differentiate finely between adjacent positions rather than lumping them together. A version of the quiz aimed at a different audience would calibrate that density differently.

4 · How we pick your mentor and kindred spirit

This is the bit that's genuinely new. Most quizzes pick a single label and stop. We pick three: an archetype, a thinker, and a fictional character, and the second two are deliberately allowed to contradict the first.

The archetype is your overall shape. But the mentor and kindred are matched independently against your specific position in nine-dimensional space, drawing from a pool of 28 thinkers and 18 fictional characters whose own coordinates have been plotted from their writings, recorded positions, or canonical fictional behaviour.

This is intentional. If you're economically communist but culturally traditional, a position that confuses two-axis quizzes, you'd get Old Socialist as your archetype, but your mentor might be R.H. Tawney (Christian socialist) rather than Marx, and your kindred might be a character with similar internal tension. The matching is designed to honour your contradictions rather than smooth them out.

5 · The Foundation Test

The three vignettes in Chapter III (the dog, the flag, the funeral joke) are adapted from Jonathan Haidt's Moral Foundations Theory (MFT), specifically the research stimuli used in his 2001 paper "The Emotional Dog and Its Rational Tail" and developed further in The Righteous Mind (2012).

MFT proposes that moral reasoning operates across six foundations: Care/Harm, Fairness/Cheating, Loyalty/Betrayal, Authority/Subversion, Sanctity/Degradation, and Liberty/Oppression. Different political tribes weight these differently. Broadly, the left prioritises Care and Fairness; the right activates more foundations including Loyalty, Authority and Sanctity. The dog scenario probes Sanctity; the flag scenario probes Loyalty; the funeral joke probes Authority.

Our adaptations score against four of our nine axes (TRA, SOC, NAT, GLO). This is the most peer-reviewed part of the quiz. The original Moral Foundations Questionnaire is available at moralfoundations.org.

6 · How we score the UK parties

We position the five parties based on what they actually do, voting records, legislation passed and public policy decisions taken in 2024 to 2026, rather than what their manifestos promise. Manifestos are aspirational fiction. Records are not.

Alignment is calculated using cosine similarity, which measures the angle between your 8D vector and each party's, not the distance. This is the right metric for "ideological direction" rather than "ideological intensity": a moderate Labour voter and an angry Labour voter point the same way, just at different magnitudes.

The parties, mapped

All five parties plotted on the same nine axes. Each axis runs from −100 (the inner end of the spectrum) to +100 (the outer pole). The shapes overlap because party positions overlap, what matters is the direction each shape points, not its size.

These coordinates will date. We'll update them as the political landscape moves: if a party shifts position significantly, the numbers will move with it.

7 · What we can't do

• It's UK-centric. The party-alignment section won't apply outside Britain. The axes themselves are roughly portable; the archetypes are calibrated to UK political traditions and would need re-calibrating for other contexts.
• The axes are necessarily simplified. "Tradition vs progress" packages dozens of distinct attitudes into one dimension. The same is true of every axis. This is a feature of dimensionality reduction, not a bug, but it does mean nuance is lost.
• The party scoring is a snapshot. Parties move. Labour in 2030 won't be Labour in 2026. We've timestamped this version (May 2026) and will update it when the picture changes.
• It can't predict your vote. Politics is downstream of identity, habit, family, geography, and a thousand other things this quiz doesn't ask about. A 70% Green alignment doesn't mean you should vote Green or even will.
• The matching algorithm has rough edges. With only 28 thinkers and 18 kindred spirits, some 8D positions are not perfectly covered. We've designed it to fail gracefully: when no good match exists, the system says so rather than fabricating one, but the further you are from a well-trodden ideological position, the looser the fit will be.