🇸🇪 Svenska | 🇬🇧 English | 🇪🇸 Español

Statistics should help us understand reality. In political debates, they are too often used to defend a conviction that has already been formed.

I have reviewed a list of names of people serving life sentences in Sweden in 2019. My estimate is that around 95 per cent were men and that around 65 per cent of the entire group had names that I did not perceive as Swedish.

Names establish neither country of birth, citizenship nor a background outside Europe. People serving life sentences are a specific group that does not represent all prisoners, but more broadly we also know that both men and people of foreign origin are overrepresented among those convicted and imprisoned.

Nevertheless, the review illustrates how easily the same material can be used to tell different political stories. Anyone who wants to make men the problem highlights sex. Anyone who wants to make immigrants the problem highlights origin. The selected characteristic is then treated as an explanation.

My argument is that feminist reasoning that places collective blame on men and racist reasoning that places collective blame on groups defined by origin can use the same mental shortcut. They seek support for their political goals rather than knowledge that could change those goals.

But a distribution is not a complete causal explanation. The fact that almost everyone in the group reviewed is male does not mean that almost all men are criminals. Nor can my assessment of names show what proportion of any particular group defined by origin commits crimes. Fewer than one in 50,000 people in Sweden are serving a life sentence, so this is clearly a vanishingly small minority among both men and immigrants.

To discuss overrepresentation, we also need a comparable population share and clear definitions. To understand the causes, we need to examine factors including age, childhood circumstances, norms and criminal environments. Sex and background can also overlap.

We must be willing to investigate uncomfortable patterns. But we must be equally willing to scrutinise our evidence and abandon our favourite explanations. That applies to my own estimates too.

Politics should follow knowledge, and measures should be judged by their results. One simple question separates the desire to learn from the desire to be right:

What facts would make you change your mind?

If the map does not match the terrain, the map is wrong. Even when that map is our own ideology.