Bradford-Hill, the lead-crime hypothesis, and sociology.

[Lago PARANOIA]

Just to recap, the lead-crime hypothesis by Rick Nevin and Kevin Drum is that the crime waves and the sudden drop in crime both in the United States and internationally from the 70s to present time was not primarily caused by poverty, abortion, crack-cocaine, television violence, degrading morals of the underclass, etc. but instead through the introduction of and elimination of aerosolized lead in gasoline.

The reason why I use this as an example is because Rick Nevin anchors his hypothesis through the Bradford-Hill criteria for determining causality through correlation. You know, to answer the glib and trite reply of 'correlation does not equal causation'. Now, the BHC is generally used in epidemiology but I don't see any reason why it can't be used in other fields. Like, say, psychology or sociology.

Not to say that it's perfect or anything. I mean, here's a summary of some of its pitfalls and limitations when used for softer sciences:

Nonetheless, the rejoinder I keep getting in RL and Internet debates when this topic is brought up and I bring up BHC is that BHC cannot be used for social sciences. I mean, just from wikipedia, I get a:

The argument proposed in this line of thought is that when considering the motives behind defining causality, the Bradford Hill criteria are important to apply to complex systems such as health sciences because they are useful in prediction models where a consequence is sought; explanation models as to why causation occurred are deduced less easily from Bradford Hill criteria as the instigation of causation, rather than the consequence, is needed for these models.

Can someone explain that part to me a bit more and why that's such a masterstroke?

[name_here]

Well, the obvious comment is that sometimes two things are correlated because they have the same cause, which is less of an issue if you're able to intentionally trigger one of them and then observe the other. For instance, the ice cream sales-murder correlation resulting from both being more common in summer.

[Lago PARANOIA]

As far as the ice cream sales/riots/summer thing goes, that would've been captured by the biological plausibility criteria. There isn't a proposed and supported biological mechanism for the consumption of ice cream to correlate with violence. Heat and aggression might, however. Not sure if that one is medically supported or if it's just intuition.

[Shiritai]

Debate over the scope of application of the criteria includes whether they can be applied to social sciences. The argument proposed in this line of thought is that when considering the motives behind defining causality, the Bradford Hill criteria are important to apply to complex systems such as health sciences because they are useful in prediction models where a consequence is sought; explanation models as to why causation occurred are deduced less easily from Bradford Hill criteria as the instigation of causation, rather than the consequence, is needed for these models.

I hate the phrasing of this section; it's ambiguous as to whether it's a critique of or support for applying Bradford Hill criteria to the social sciences. The only somewhat sensible thing it says is that the Bradford Hill criteria are more useful for finding the triggers for events, and not as useful for finding the mechanism of action. But that's pretty damn obvious, since they're not made for finding explanations. I'm as confused as you as to why this means it's unhelpful in social sciences, because I'd thinksocial scientists would still want to know whether there's causality between events even if they can't explain the "why" immediately. But maybe the people who disagree are of the Austrian school; who knows?

[fectin]

BHC are not actually a checklist of criteria. They're a series of prompts to think more deeply about your apparent correlations.

Of course, the same is true of statistical significance, yet here we are 90 years later and that's become a (completely arbitrary) gold standard.

[Occluded Sun]

People are opposed to rigorous standards of reasoning in the 'softer' sciences because the widespread adoption and use of such standards would interfere in their ability to peddle nonsense at will.

If they have an agenda - almost any agenda - they will be opposed to skepticism and rigorous thought. They don't mind teaching people look critically at other parties' propaganda, but teaching children to critique propaganda itself undercuts their own influence. Every group wants the public to be resistant to their enemies' propaganda, but vulnerable to their own.

It's a large part of why the perennial attempts to teach skepticism as part of school curricula never catch on.

[tussock]

Was there even a crime wave? I recall that sentences for drug abuse went from basically sleeping it off to life in prison in a bunch of places, which obviously increases "crime" without really changing anything. The murder rate in some places did go up, but there was significant police corruption associated that later got dealt with to some extent.

Western police corruption was fucking awful in the 70's. I mean, there's places now have huge murder rates and also have massive police corruption, but no lead in their petrol.

But yeh, if there was a proper crime hump in places where leaded petrol use happened, and it's strongly associated with concentrated use of that petrol around the world, then that's a pretty good argument. Because the mechanism for action can just be lead damage to your brain, which you can't actually test or prove in any way other than by correlation.


So, as demonstrated in my own post, "not using BHC" is about not using it blindly and making sure the study accounts for the myriad of other possible social factors. Which I imagine it did, but I clearly haven't read it, so questions abound.

[Lago PARANOIA]

Was there even a crime wave? I recall that sentences for drug abuse went from basically sleeping it off to life in prison in a bunch of places, which obviously increases "crime" without really changing anything. The murder rate in some places did go up, but there was significant police corruption associated that later got dealt with to some extent.

The crime waves and subsequent decline in crime were international in places that provided leaded gasoline -- even in countries that didn't have as extreme of a response to the crime waves/war on drugs as the United States. Even if you just restrict the analysis to the United States, the rapid rise and decline in, say, murders in Chicago, Los Angeles, and New York City were very real. You can listen to people go on all day about just what constitutes a burglary or sexual assault or whatever, but a dead body is a dead body. And as much as I hate the United States' frankly racist overreaction and misdiagnosis, the crime epidemic did deserve a response.

Of course, one of the implications of the lead-crime hypothesis is that as much as the Guilianis and O'Malleys say otherwise, changes in policing and sentencing policy and soforth are mostly self-aggrandizing bullshit. Which is what you'd expect, but, you know how people are.

[tussock]

Thanks Lago. I've seen a few reports about how policing changes follow crime rates rather than leading them. One of the reasons so many US cities got "tough on crime" in the 90's was the crime had reduced and officers had a bit more time on their hands for the small stuff. A few years later, claim victory. Woot!

Funny that it likely reduced because the US banned lead paint in the 70's and phased out leaded petrol from the same time. Funnily enough it'll be hard to check for all countries because lead paint was banned in many by 1920. Australia issued caution that lead paint not be used in places where children might play in 1904.

Heh, these days, huge government cautions about dealing with old lead paint, don't sand it, don't leave any chips. Eh, much the same way Asbestos or 245T or all sorts got dealt with, so long as it was profitable with no ready economic alternative you just don't worry.

Edit: Hah, yep, same lead additive still used in Avgas, because the alternatives are "too expensive". Live near a busy airport, anyone? Might want to move before you have kids.