Scott Alexander

Adult accident death rates have also gone way down, and adult murder rates stayed about the same since the 1960s.

The simplest explanation is that child trends simply mirror adults, and the adult trend certainly isn't caused by coddling.

So maybe the kids aren't either.

What about kidnapping?

Plenty of people are kidnapped, but it's usually something like a relative stealing them away in defiance of a divorce custody agreement.

The traditional kidnapping, where a creep in a white van plucks a child off the streets, is much rarer.

Only about a hundred such incidents come to the attention of the authorities per year.

Even if the true number is ten times higher, is this a reasonable multiplier?

That's still only one in 70,000 children per year.

On the other hand, over around 10 years of relevant childhood, I'm assuming babies can't get into white vans and nobody wants to kidnap 17-year-olds, there's a 1 in 7,000 chance.

And surely, if you're the only person still letting your kid play outside, your chances are worse than average.

But if you're an upper-class person in a good neighborhood, your chances get better than average again.

Let's estimate a 10 times penalty for playing outside and a 2 times bonus for not being a poor person in a ghetto.

Now it seems like your total chance of child abduction per child, lifetime, is 1 in 1,400.

but it seems like most abductions are navigated successfully.

Quote, Quote,

End quote.

Scott writes, if 4 in 180 abducted children are killed, I don't know if the number of Amber Alerts is the right denominator, that's a 1 in 63,000 chance of child death via abduction per child lifetime.

Seems not great, but not like a deal breaker.

So maybe the more relevant question is, if you decide you want to do this, is it even possible?