Yes, that's true. Political thought is highly qualitative and can be measured across many other dimensions, such as Globalism vs. Nationalism or Traditionalist vs. Progressive. While other tests explore these many facets to provide a more granular and precise metric, this test is intentionally designed for simplicity and clarity. When you condense a user's political leanings to a simple left-right scale, you necessarily lose some specificity. For example, if two people score a +5.00 (Moderate Right) on the social scale, it does not necessarily mean they will have identical views. This is just a fundmanetal tradeoff. Generality and interpretability necessarily comes at the expense of specificity.
The reason this test only has two axes is because, whenever people generally discuss their political leanings, they typically categorize themselves in two primary ways: "I am socially [X] and economically [Y]". The news often summarizes movements and positions by politicians using similar language(eg. "Keir Starmer has drifted to the right since the rise of the far right in Britain." , "Mark Carney is generally viewed as more economically right-wing than Justin Trudeau.") By focusing on these two fundamental pillars, this test provides a result that is immediately interpretable and easy to communicate. This two-axis approach ensures that your results remain clear, useful, and communicable for broad comparisons.
The aim of the test is to allow the user to ponder and form opinions on tough and relevant propositions regarding modern social issues. Removing the neutral option, encourages people to lean in a direction, as opposed to staying on the fence, which ultimately furthers the main goal. Moreover, if a user uses the neutral option too many times, the final score will necessarily become meaningless.
The guiding principle was modernity. What is being actively discussed by politicians and political candidates in the western world. Immigration, environmental issues, LGBTQ issues, income inequality are all extremely common themes in political parlance and are topics which are hotly contested, so there were many questions associated with these topics.
In other tests there are often outlandish or outdated questions that a near complete majority of the population would have an identical position on. With questions that are not controversial in modern times, it will necessarily lead to a "clumping" of political scores. This would make it difficult to distinguish political leanings amongst peers.
By choosing highly relevant questions and topics, this test is designed to avoid those dead zones, making it easier to distinguish meaningful differences in political leanings among your peers.
Yes, they are, because it is impossible to frame a question in a completely unbiased manner. Every possible variant of a question can bias an individual in a particular way.
Let's take the question: Gender transitioning care should be permissible for minors. This question can be framed in many ways:
Each of these question variants can bias an individual in varied ways, so coming up with one that has zero bias whatsoever is impossible.
However, in the test I did make a sustained effort in keeping bias to a minumum. Approximately half the questions are framed where Strongly Agree is linked to right-leaning positions, and the other half towards left-leaning positions. For the sake of consistency, for most of the questions, I avoided questions with negation vocabulary such as "not", "un-", or "dis-."
Yes, it's nearly impossible to have a global left and global right that fits into the a vast majority of countries. Developed countries and developing countries fundamentally face different challenges and have a vast array of governance styles. The political parlance of developing countries will necessarily be vastly different, ergo the political left and right will also be vastly different.
To provide a consistent and meaningful benchmark, I chose to focus on the common issues currently facing major Western countries. It allows for a more accurate approximation of political tensions within the contexts where these specific discourse topics—such as immigration, environment, and social justice—are most actively debated and polarized.
These questions are propositions and sometimes deliberately vague. They're meant to be answered more from the emotional responses and initial reactions of the user than detailed opinions on the specific policy.
Let's take the question: Inheritance tax should be low.
It's completely fair to wonder what is meant by "low." Is it 5%, 15%, 25%? Is the inheritance tax referring to people of high income or low income? Inheritance taxes often differ for different income brackets, so which group is facing the low inheritance tax?
The user can go on and on pondering the specificities of the question. But fundamentally, the question should be answered through the initial reaction or emotional response of the user when faced with the proposition.
Someone who is more left-leaning is --likely-- to say, "Rich people passing down their inheritances is the primary reason income inequality has been exacerbated in recent decades. Inheritance tax should be high, so that income can be redistributed and prevent the creation of a permanent hereditary elite."
Someone who is more right-leaning is --likely-- to say, "I worked hard to build the wealth I made, people who are wealthy worked hard to create the wealth they made, and they should have the right to pass it on to their children. Why should the government effectively tax my income and wealth twice? Inheritance tax should be low or non-existent."
Yes, they do. Political thought is highly qualitative by nature, so it's very difficult to create a test that is both comprehensive and perfectly partitioned.
This is a common and valid observation. Political norms vary significantly by nation; for example, government provided healthcare might be a polarized "left vs. right" issue in the US, but enjoys nearly universal support in a country like the UK or France. In this case, someone might feel that supporting a universal government funded healthcare doesn't make them "right-wing" within their own country's context.
However, to provide results that are comparable across countries, it is necessary to establish relatively fixed definitions and a universal benchmark. This allows us to account for "country-fixed effects"—the reality that the average political position in one country, like the UK, may be naturally further to the left than that of another, like the US. By using a standardized scale, the test measures where you stand on a global spectrum rather than adjusting for the specific cultural shifts of every individual nation.
They're generated by prompting AI to respond to the questions from the point of view of the individual. However, AI is imperfect and different AIs generate different answers, so these scores are an approximation. The major issue with the AI generated opinions it that AI "figures out" that the questions are part of a political test and makes inferences for positions as opposed to strictly researching opinions. For example, Kemi Badenoch is the right-leaning PM of the UK and AI argued that she would Strongly Disaree or Disagree to the question "Healthcare should be primarily funded and run by the government." Given that agreeing with this position is right-leaning, AI infers that she would Disagree with this position. However, the UK has the NHS, a publicly funded healthcare system, that is free at the point of use for all UK nationals. Badenoch nor the Conservative Party have ever advocated privatising healthcare or getting rid of the NHS, so in fact Badenoch would likely Agree or Strongly Agree with this question. The best way to root out these inaccuracies is to research the positions for each question one-by-one for each and every politician to guage what they would their position would be for a particular prompt. I'll likely do that when I have more time.
It's also worth mentioning that there will naturally be differences between political candidates and elected politicians. After entering office, politicians often moderate their positions to be more pallatable to the general popualtion. For example, Giorgia Meloni (PM of Italy) was often labelled to be a "far-right" politician on the campaign trail due to her espousing strong euroskeptic and anti-migrant views. However, after entering office she has drastically moderated her tune by cooperating significantly with the EU and authorized many non-EU work visas. Naturally, when the scores are being generated, it will assign positions based off their track record in office, and hence some politicians will have more moderate scores than expected.
Politicians also are often dishonest, so they may make comments on the campaign trail that are different from the action that they actually take.
This is a toss-up choice based on varying global conventions. In most of the world—including the UK, Canada, France, and Germany—red is traditionally associated with left-leaning parties and blue with the right. However, in the United States, this convention is reversed: red is synonymous with Republican conservatism, while blue represents Democratic liberalism.
I wanted to keep the test at 50 questions exactly and avoid too much overlap with the existing questions. But these are some of the questions that were not included, if you're curious.