Answers and results are calculated in your browser.

Methodology

How the source questions, paired answers, twelve axes, display projection and ideology comparisons work.

Sources and original content

This independent educational questionnaire adapts the 12axes source in politicaltests/politicaltests.github.io at commit 5a508de557cc88eb7e9a44d89f45501dbe912d61. It is not affiliated with politicalcompass.org. The original bank contains 288 entries across twelve axes, with 287 distinct strings. One hundred verbatim anchors are selected from that bank. The open-source license preserves the upstream MIT notice; every original anchor retains a file, axis, group and item reference.

Each anchor has one newly written reverse statement, labeled site-original in questions-pairs.ts. The new wording is separate editorial content outside the upstream MIT notice, not a validated psychometric instrument. Quick, Standard and Deep contain nested sets of 15, 30 and 100 pairs: 30, 60 and 200 statements. A pair has identical wording in every length. Every axis has at least two pairs in Standard and eight in Deep. Economic and social dimensions receive the additional coverage. A seeded shuffle changes presentation order only; a full block separates the two wordings.

The twelve original scales

The seven original response values remain +2, +1, +½, 0, −½, −1 and −2. Original source groups 0–3 count positively; groups 4–7 count negatively. Each reverse statement uses the opposite coding and contributes to the same axis. There is no extra weighting for stronger-sounding source groups. For each axis, sum response times direction and normalize by the number of its answered statements. Neutral responses produce 50. The two poles sum to 100; these are balances of answers, not population percentiles.

The formula retains the existing generalized normalization: P = 50 + 25 × sum(response × direction) / statement count. For the original short quiz's five questions per axis, this agrees with its 50 + 5 × sum rule. Internal calculation keeps the existing rounding and precision. Visible percentages and coordinates are rounded to integers only when displayed; matching uses the underlying scores. Short versions have fewer observations and do not justify fine distinctions.

Economic axis: the horizontal projection

The chart's X axis combines economic ownership (f) and economic control (j). A stronger collective-ownership or planning emphasis points left; private ownership and market allocation point right. Trade is measured separately and is not an input to this display projection. The bar labeled Economic axis is the original ownership scale; the horizontal coordinate is the stated composite.

For projection version economic-social-v1, the ownership and planning weights are 50% and 50%. These are editorial estimates, not empirically calibrated weights. The combined score E = (50 × f + 50 × j) / 100 is centered and scaled as X = (50 − E) / 5. This produces a range from −10 on the left to +10 on the right. The projection does not change any of the original twelve scores.

Social axis: the vertical projection

The Y axis combines state security versus freedom (e) with democratic representation (b). Representation is inverted before combining: stronger security powers and less democratic representation point upward, toward authoritarian government. Greater personal freedom and democratic participation point downward, toward libertarian government. Cultural change and religion remain separate dimensions; neither is treated as a synonym for personal liberty.

The security and inverted representation weights are 50% and 50%, again editorial estimates without a validation study. S = (50 × e + 50 × (100 − b)) / 100, then Y = (S − 50) / 5. The range is −10 to +10. Equal midpoint scores place the dot at the center. The website and locally generated share card use the same projection function and version.

Paired-answer consistency score

This separate feedback compares two completed responses after both are coded onto the same axis direction. Average their absolute difference, divide by the maximum response difference of four, subtract from one, and multiply by 100. Round the final result to an integer. Higher scores mean greater agreement within these pairs; the score never changes political scores, question weights or ideology ranking. Fewer than five complete pairs produces no score or grade, rather than a default zero.

Neutral responses to both statements count as consistent, so a high score cannot establish knowledge, care, truth or validity. Differences can reflect context, ambiguity or changing judgment; they do not prove that wording caused an effect. The political compass test offers this as a reflection aid. Only the twelve political scores enter a share link, so the consistency feedback cannot be recovered from a shared summary.

Inherited ideology matching and limits

All 38 upstream candidates, coordinates and original tie order remain intact. Matching uses economic ownership f, security e, trade/protectionism k and cultural change h. The inherited diplomatic coordinate is mapped to k, not the separate international-relations scale c. Squared Euclidean distance determines order: D = (f − econ)² + (e − govt)² + (k − dipl)² + (h − scty)². The displayed integer similarity is 100 × (1 − square root of D / 200). It is a distance index, not a probability, and the three percentages need not sum to 100.

The display chart is a different projection and is never substituted into the four-dimensional matcher. Unpublished ideology articles are represented by plain-text names. Country, party and historical profiles are outside this release's reference display: the old single-source records do not meet the revised evidence threshold. Future country and party references require two public sources per object. Historical comparisons will include deceased people only and use five proximity bands rather than percentages. No claim of an accurate or precise match is made.

Offline operation and local information

After the first screen renders, the Service Worker downloads this release's public static resources. Offline readiness appears only after the cache is complete and the page is controlled. Answering, backtracking and results use states in the same page and do not change its URL. Requests fulfilled by the Service Worker use the local cache; this is distinct from a network transfer. First downloads, updates and deliberately opening another website are separate operations.

Read the privacy policy for browser storage, cache clearing, optional sharing and what happens when storage is unavailable. No free answers or scores are uploaded. This educational model has no claimed scientific validation, and inherited questions may be loaded or combine multiple issues. It offers a snapshot of answers, not a diagnosis, prediction or fixed identity.