I built a political compass alternative with 5 axes, 80 calibrated questions, and results compared against 50 countries and 24 world leaders — feedback welcome

I've been frustrated with the standard Political Compass test for months since I first took it back in January. Researchers have documented that it systematically pushes respondents toward the libertarian-left, regardless of their actual views, and its questions haven't been updated since 2001. Most people's results cluster in one corner of a four-quadrant grid — which tells you almost nothing.

As a result, I built my own from scratch. Here's everything to know:

The test itself:

— There are 5 independent axes instead of 2: Economic, Social, Civil Liberties, Governance, and System Trust. No other political test I've found measures that.

— 3 difficulty modes (40, 50, or 80 questions). Additionally, everything is mathematically balanced, which you can verify by reading the source.

— Every single question has an optional context button. When you click it, you get further context on the question if you are struggling to understand it with just the original premise.

— Questions are about things actually happening right now: NATO funding, encryption rights, trade tariffs, police reform, platform moderation. Specific on purpose.

What is shown in the results:

— 14 named, defined ideologies with full plain-English explanations

— A comparison against 24 political figures (I personally researched and placed every one of them across all 5 axes)

— A comparison against 50 countries (same thing — all placed by me)

— A shareable result card you can download as an image

— URL sharing so you can send your exact result to someone

The research and calibration:

I spent an immense amount of time researching the political science literature on how ideological axes should be defined and separated before writing a single question. I read academic critiques of the original Political Compass — there are serious ones, published in peer-reviewed political science journals — and used them to inform how I structured the axes and balanced the question set.

I then ran the test with hundreds of real people to calibrate it. Additionally, as I am currently in a semester-long program that focuses on politics, I had students who are very well-informed about their own politics take it to help figure out whether there were any glitches that could lead to incorrect results, as they all know their own views extremely well. I collected results, identified where the scoring felt off, and reworked or reweighted the questions that produced skewed distributions. The ideology centroids (the mathematical coordinates that define where each of the 14 ideologies sits in 5-dimensional space) were iteratively refined based on real response data. That calibration process is a big part of why I think the results feel accurate to people, which is the hardest and most important thing to get right in something like this.

On the AI question:

I used Claude to help build the interface and frontend code. I want to be upfront about that. But every single question was written by me. Every definition of an ideology was written by me. The placement of all 24 political figures and 50 countries across 5 axes was researched and decided by me. The context blurbs on every question, the descriptions of what each result means, the "why this test exists" writeup were also all done by me. The AI helped me build the thing I wanted to build. The substance of what the test actually is mine.

This has been a long-running project, and honestly, the thing I'm most proud of making. I'd love feedback, especially if you think a question is biased, or your result felt wrong.

findmypolitics.com — free, anonymous, no account needed 🙂