What a country average is

A country average on Uncanted is the mean position of finished quizzes from accounts that chose that country. Economics is averaged. Authority is averaged. The result is one dot. It is drawn on the world-leaders chart and, for the United States, on the United States chart as well. It is not a census. It is not a poll. It is not weighted by age, race, state, or likelihood of voting. It is the people who made an account, picked a country, and finished the test.

Only the first finished quiz on an account is included. A second quiz on the same account does not move the dot. That rule is there because a mean is easy to stuff. If later quizzes counted, someone could add tests and tug the country toward themselves. The first quiz is the one that counts. Later quizzes are for the account, not for the map. An unfinished quiz does not count either. A country has to be saved. An account with no country is absent, not assigned to the United States by default.

The average appears in full only after that view is opened. The charts can say what the dot represents. They do not publish the coordinates in the public text. A preview of the number would make the sentence “this is the average” into the result itself. The description is public. The coordinate is not, until the view is opened. Nothing about that changes the math.

Small numbers are noisy. A country with four finished quizzes will jump when the fifth person finishes. The tally page, which is not public, exists so the site’s owner can see those counts. A reader looking at a green dot should assume the sample might be thin, especially outside the United States, and should read the sentence that says it is not a national poll. A dot with a large American file is still not a national poll. It is a large file of self-selected test takers. Self-selection is not fixed by volume.

The average can sit somewhere no single person sits. That is what a mean does. If half the file is left and libertarian and half is right and authoritarian, the mean can land near the center, a place neither half chose. The chart does not draw the cloud, only the center. A reader who matches the average may match a compromise that nobody in the file actually holds. The results page does not call that person “the average American” as a type. The label on the chart is a statistic.

Comparing a reader to the average is optional and easy to overread. Being far from the American dot means being far from this site’s American file. It might mean the reader is far from the country. It might mean the people who finish a thirty-question political test are a strange sample, and the reader is strange in a different way. The site cannot tell those apart. It does not try. The sentence under the dot is the whole of the claim: mean of first finished quizzes, accounts that chose this country, not a poll.

Other countries are computed the same way. A dot for Canada is not a statement about Parliament. It is a statement about accounts that selected Canada. If nobody has finished, there is no dot. The site does not borrow a national stereotype to fill the gap. An empty country is better than a fake one. When a dot does appear, it can move as new first quizzes arrive. It should not move because someone retook the test.

Use the average as a landmark, the way you would use a mile marker you know was placed by the people who happened to walk that road. It tells you where this file has gathered. It does not tell you where a country “really” is. Anyone who wants that question needs a poll with a sampling frame. Uncanted is not that poll, and the average page is not going to become one by adding an adjective.

All articles · Methodology