What a typology of profiles can say, and what it invents

A typology starts from positions measured on several dimensions and groups together people who resemble each other, generally by attaching them to the profile they are closest to. The profile is therefore not a category discovered in nature: it is a point of reference, and membership is a question of distance.

This mechanism has a direct consequence. Two people attached to the same profile may be quite far from each other, and a person located between two profiles is attached to one of the two by a sometimes minuscule difference. The name of the profile is a practical summary, not an identity.

The number of profiles is also a choice. Too few, and each group becomes heterogeneous; too many, and neighbouring profiles cease to be distinguishable and attachment becomes unstable. No number is objectively correct, which should make anyone claiming otherwise modest.

The invention begins when the name of the profile starts explaining something other than the answers that produced it. A profile may legitimately say: on these dimensions, your positions resemble these. It cannot say why you hold them, what you will do, or what personal qualities you would have.

The second slippage is the flattering description. A sufficiently general text seems accurate to almost everyone, a phenomenon known as the Barnum effect, brought to light by the psychologist Bertram Forer. It is a real risk for all profile tests, and the only remedy is to say things precise enough that they could be wrong.

The third slippage is the subtlest: the profile becomes a self-explanation that one adopts. Once one thinks of oneself as contestatory or pragmatic, one interprets one's own reactions through that word, which produces a worthless confirmation since it comes from the name, not from the facts.

Three indications are enough, and a serious model should provide them. The distance between your position and the profile retained. The distance to the nearest neighbouring profile. And the variation observed when the same person retakes the test a few weeks apart.

A tight attachment means the name retained is the least bad summary, not a clear-cut conclusion. That is useful information, and hiding it to give an impression of precision is a form of gentle deception, very widespread.

It would be excessive to conclude that a typology is of no use. It makes comparable what was not, allows a person to be situated among others, and gives a common language for speaking of configurations that recur.

It also has a pedagogical virtue: a name holds attention where a series of scores says nothing. The problem is not the name, it is using it as though the frontiers existed.

Look first at the positions by dimension, which are the measurement, and then at the name of the profile, which is the summary. If the summary seems right to you but two positions seem wrong, it is the positions that should interest you: they refer to precise questions you can replay.

And beware of comfort. A profile that pleases you entirely, with no line that makes you wince, is probably written too broadly to say anything.