Hacker News Reader

Comment by fjsolwmv | Hacker News Reader

That's a stretch. It's just as easy to say that national instability leads to both civil war and lack of modern technology like unleqded gasoline.

View on HN

Yeah, which is why the data is backed up by multiple panel analyses that actually make the argument for causation. The US data shows that crime rates rose like clockwork approximately ~20 years after leaded gasoline is used en masse, and drops ~20 years after it was banned.

Different states introduced and subsequently banned leaded gasoline on different schedules, so this was a 'natural' experiment, unless you can think of some other causative factor that precisely explains the timing of the rise and fall of crime.

I don't understand this content-free 'correlation does not equal causation' argument.

Sure, in the absence of other evidence and a mechanism of action, we should have healthy skepticism, but it's known how lead affects brain development, particularly how it screws up development of the frontal cortex, lowering both intelligence and social inhibition.

Re natural experiment: what about the water pipes? Were they exchanged as well? I don’t think you can see this as natural experiment unless other factors are equal

TFA discusses a natural experiment with water pipes. Cities near lead production centers tend to have lead pipes, so they compared cities with that characteristic and acidic water (which leaches lead into the water) and such cities without acidic water. The study found a strong effect.

Did they also account for changes in how crime rates are reported over time?

One would have to account for hundreds of different variables before you could ever conclusively prove a connection between lead and crime.

It would be impossible to 100% conclusively and with absolute precision determine the exact connection between the two. That doesn't mean it's not the best available theory at the moment, that fits the most data. There are international, interstate, and even local level analysis all pointing to the same thing. There are longitudinal studies on children pointing to the same thing. And there are established and direct casual mechanisms in the brain biology that explain it. So what does waving a hand and saying "Nah nah can't prove it" do? By your standards do we have zero knowledge on all sociological matters, since it's impossible to design and implement a perfectly controlled study?

>Did they also account for changes in how crime rates are reported over time?

The Bureau of Justice Statistics conducts a yearly survey asking people about the crimes they were a victim of in the past year regardless if the crimes were reported. The Bureau of Justice survey shows a much more dramatic reduction in crime than if you just look at reported crimes.

http://www.pewresearch.org/fact-tank/2018/01/30/5-facts-abou...

>Using the FBI numbers, the violent crime rate fell 48% between 1993 and 2016. Using the BJS data, the rate fell 74% during that span.

Reply on news.ycombinator.com