What it means to be rare in an opinion profile

A position is rare if few people stand at the same point on the same dimension. That makes it neither better, nor more considered, nor more original in the intellectual sense: on a graded scale, extreme positions are mechanically less populated than middling ones, as on any measurement.

From this follows a counter-intuitive consequence. Answering in a very clear-cut way almost always produces a rarer profile than answering in a nuanced way, without the first attitude being better informed than the second.

That is also what makes rarity commercially seductive and methodologically suspect: it flatters, it sticks in the memory, and it is obtained without teaching anyone anything.

A percentage presupposes a reference set. The people who have taken the same test do not constitute a representative sample of the general population, since they came of their own accord, by particular routes, and often because the subject interests them.

A rarity calculated on that basis therefore describes a position within a group of visitors, not within a country. An honest model must say so, and avoid the wording that slides from one to the other.

The nuance is not cosmetic. On a political test, the most motivated visitors often hold more clear-cut positions than the average, which makes moderate positions rarer in the sample than they are in the population. Rarity may therefore reverse depending on the reference chosen.

Two things often conflated must be distinguished. An isolated position may be common, while the combination of several common positions is rare: that is even the most frequent case as soon as one multiplies dimensions, since the number of possible combinations grows very fast.

It is this rarity of combination that is the most interesting to comment on, because it describes an arrangement and not merely an intensity. It also calls for more caution: a small number of different answers is enough to make it vary sharply.

Three precautions distinguish an honest use from a sales argument. Naming the reference population. Not announcing a percentage when the position is close to the centre, where the gaps are too small to mean anything. And never presenting rarity as a personal distinction.

An acceptable wording looks like this: among the people who have taken this test, few combine these positions. A misleading wording looks like this: you are among the three per cent rarest.

Rarity is a point of context, not a result. What really concerns you are your positions by dimension, and above all those where you are furthest from what you imagined.

If a test insists a great deal on your singularity and little on the content of your positions, the order of its priorities tells you more about it than about you.