Understanding convictions, your own included

A Dutch study followed the same people for two decades. Its main result contradicts the intuition of a stable circle of friends.

The episode is frequent, and so is the discomfort it produces. What that discomfort reveals has less to do with tears than with a boundary that work has moved without saying so.

They are reputed to be doomed. The available work does not confirm it, and the critical moment is not the one people think.

A rarity percentage impresses, but it means nothing until one has specified rare compared with what, and measured how.

Heritability is one of the most misunderstood words in statistics. It does not describe an individual, and its value changes with the population in which it is computed.

The published number comes from claims recognised by the insurance scheme. What is not reported appears nowhere, and that share is estimated by other means.

One is felt, the other counted. Their correlation is moderate, which is enough to invalidate most of the shortcuts used in public debate.

Understanding what the other is going through and feeling it with them are two distinct operations. One supports action, the other exhausts.

Almost every couple declares itself in favour, and the timetables tell another story. The gap is not explained by hypocrisy.

Sweden legislated in 1979, France in 2019. More than sixty states have followed, and evaluating these laws poses particular methodological difficulties.

Changing occupation may be a promotion, a fallback or a constraint. Available statistics separate those three cases poorly.

Mobile phone data made it possible to trace a curve that self-report surveys suggested without demonstrating. Its shape is more regular than expected.

Those who show nothing are credited with feeling nothing. The physiological measurements often say the opposite.

We reread the first weeks the way we reread a detective novel, looking for clues. Most of them were not clues.

A poll is a measuring instrument with a known precision and known blind spots. The problem rarely comes from the instrument, often from how it is read.