Our Methodology

We publish our scoring method because a test that hides how it works cannot be checked, and a political instrument that cannot be checked should not be trusted.

The two-axis structure descends from Hans Eysenck's factor-analytic work in the 1950s. Applying factor analysis to survey responses on political attitudes, Eysenck found that two underlying dimensions were needed to account for the observed pattern of correlations: a familiar economic dimension, and a second, independent dimension tracking attitudes to authority, discipline and conformity. Crucially, that second dimension appeared at both economic extremes. The visual format of the compass owes more to David Nolan's 1971 chart, which arranged economic and personal freedom as two axes of a square.

Each of the fifteen statements is assigned to one axis and carries a weight. The weight has a sign, which indicates the direction in which agreement moves the score, and a magnitude, which reflects how strongly responses to that statement discriminate between positions. A statement about progressive taxation carries its weight on the economic axis; a statement about state surveillance carries its weight on the social axis.

Scoring proceeds as follows. Your answer on the five-point scale is converted to a value from -2 to +2. Each value is multiplied by the weight of its statement. The products are summed separately for each axis. Each sum is then divided by the maximum score attainable on that axis and multiplied by 10, producing a final coordinate between -10 and +10. Because normalisation uses the maximum attainable rather than the observed range, the neutral midpoint of the answer scale corresponds to the centre of the compass. This matters: several well-known public compass tests place a large majority of respondents in one quadrant, which indicates calibration rather than a discovery about the population.

The statements are distributed evenly between the two axes so that neither is measured more precisely than the other. They are worded to probe underlying principles rather than named policies or parties, because party positions shift and a question tied to one becomes uninformative within a few years.

We review the question set periodically. The purpose of review is to check that each statement still discriminates between positions as intended. Political language changes, and a statement that divided opinion informatively a decade ago can become either universally accepted or unintelligible. Where a statement stops doing useful work it is replaced rather than retained for continuity.

Finally, a caution about all instruments of this kind, ours included. Where the axes are placed, how the statements are worded and which positions end up near the centre are choices made by whoever built the test, and different defensible choices produce different results for the same respondent. A compass result is a prompt for reflection, not a measurement in the sense that a thermometer produces one.