Merged: Lead (Pb) exposure linked to crime rates

[Dinwar]

Quote:

which I do not understand. Maybe this addresses part of your question regarding other social parameters (like abortion)? I honestly have no freaking idea. Perhaps if you read the paper, you would understand, and you could explain to me if that is the case.

I don't get it either. Seems like a lot of weasle words. "Elasticity" isn't a common term in any of the statistical methodologies I'm familiar with, though I'll admit that I'm no statistician.

I also am strongly biased against inferring causal relationships from mere correlation. I always want to see the actual mechanism. If you say "X causes 56% of Y" I want to see what the mechanism is, why 44% of the time the mechanism breaks down, etc.

I will say that the data seem to support the notion that poor impuls control or other neurological factors are playing a role in crime rates (though I think it would be best to figure out why New York and Washington, D.C. don't follow). As I understand it most violent crime is committed by first-time offenders--ie, someone gets extremely angry and shoots someone/stabs them/bashes them over the head. Most crime isn't commited by hardened, systematic criminals (at least no PROSECUTED crime, which is another significant factor when dealing with crime rates). That would imply that it's impulse control. Property crime is a different set of activities, which may or may not be related to impulse control and almost certainly aren't related to agression control. No one ever says "I'm so mad, I could spraypaint a fence!"

That said, again, there were a lot of changes going on at that time. Changes in diet, changes in education, etc. It could easily be that those areas that called for abrupt reductions in lead use embraced more social changes rapidly as well, while those which called for a gradual phase-out of lead use also instituted social changes more gradually.

Originally Posted by Kestrel

Like you, I don't see this as proof that we have found the root source of all crime. But it is an interesting theory that merits further study. There is a plausible mechanism and not just a correlation. Childhood lead exposure could be one of the factors leading to criminal behavior.

I'm willing to go so far as to say it's interesting and warrants further study. I'd recommend looking at Africa, South America, and other non-Western countries to see if there are similar trends (ie, if different societies react differently to the same forcing mechanism). If they don't, it's an indication (not strong, but as strong as the original data) that there's more to this story than just lead causing problems. Even there, though, you run into the issue of how to disentangle the signal caused by lead from the signal caused by the rest of the social changes inevitably going on (no one ever just enacts environmental change--it's always in the middle of other social changes).

Originally Posted by rjh01

Also stop saying we should deal with certain people who commit crimes harshly in order to deter others. This would not work.

Doesn't follow from the article. The issue is, not everyone who grew up in neighborhoods with significant lead levels becomes a criminal. Obviously lead isn't the only issue here. It may be that harsh punishment will yield results--impulse control can be learned, after all (that's what childhood IS, learning to control your impulses). I'm not saying it's a good method; I'm just saying that this article says nothing about punishment for crime, nor does it really address the roots of crime. It merely discusses one contributing factor.

[bruto]

Originally Posted by Madalch

Probably not really relevant- water flowing through lead pipes will not pick up much lead. Acidic liquids (particularly ones full of organic ligands such as tartrates, including wine and cider) will pick up lead much more quickly.

Indeed, although at least some Romans were aware of lead poisoning, it appears that it did not occur much from the water pipes, in which the continuous flow of lime-heavy water combined for little risk. Unfortunately, many Romans used lead for cooking pots and vessels, and as a preservative for wine and other substances. Amusiningly, Frontinus, whose fascinating and detailed account of the water system I have read, is said to have died himself of lead poisoning, but he probably drank too much wine, not too much water.

[rjh01]

Originally Posted by rjh01

<snip> Also stop saying we should deal with certain people who commit crimes harshly in order to deter others. This would not work.

Originally Posted by Dinwar

<snip>

Doesn't follow from the article. The issue is, not everyone who grew up in neighborhoods with significant lead levels becomes a criminal. Obviously lead isn't the only issue here. It may be that harsh punishment will yield results--impulse control can be learned, after all (that's what childhood IS, learning to control your impulses). I'm not saying it's a good method; I'm just saying that this article says nothing about punishment for crime, nor does it really address the roots of crime. It merely discusses one contributing factor.

I agree that the majority of people who are exposed to lead will not get a criminal conviction. However a minority of people who are exposed to lead will get a criminal conviction. This minority of people will form a significant number (majority?) of such people. So if there are other factors they are a constant and not really relevant. Like why do these people get a criminal conviction and not others who are exposed to the same amount of lead.

I am just discussing what the implications of the article are. Because I think they are huge.

[Kuko 4000]

Originally Posted by rjh01

I am just discussing what the implications of the article are. Because I think they are huge.

rjh, you might be interested in Tom Clark's ideas about how scientific naturalism can be applied to our judicial systems and everyday thinking:

http://www.naturalism.org/criminal.htm

[St.Michael]

Is there any data avilable for Japan or other non-western countries showing a similar link?
This is interesting but could it be too good to be true, without considering other factors at play.

[casebro]

Originally Posted by Dinwar

I will say that the data seem to support the notion that poor impuls control or other neurological factors are playing a role in crime rates (though I think it would be best to figure out why New York and Washington, D.C. don't follow). As I understand it most violent crime is committed by first-time offenders--ie, someone gets extremely angry and shoots someone/stabs them/bashes them over the head. ....

There are a couple interesting tid bits in there Dinwar. I take it crime rates dropped in correlation with lead, but NOT in NYC and DC? ]

And the impulsive violent crimes- does that hold for street crimes too? Muggings, gang related, drug related.... I don't think that adds up to "most" being impulse. But it is an interesting classification, I've never heard of "impulsive violence" in the National Crime Database.

[meg]

Originally Posted by Dinwar

I will say that the data seem to support the notion that poor impuls control or other neurological factors are playing a role in crime rates (though I think it would be best to figure out why New York and Washington, D.C. don't follow). As I understand it most violent crime is committed by first-time offenders--ie, someone gets extremely angry and shoots someone/stabs them/bashes them over the head. Most crime isn't commited by hardened, systematic criminals (at least no PROSECUTED crime, which is another significant factor when dealing with crime rates). That would imply that it's impulse control. Property crime is a different set of activities, which may or may not be related to impulse control and almost certainly aren't related to agression control. No one ever says "I'm so mad, I could spraypaint a fence!"

Actually, if you look at the definition of property crime, I could see how it might be related to impulse control. According to the paper..

Quote:

Violent crime consists of robbery, murder, aggravated assault, and rape. Violent crime is dominated by assault and robbery; murders represent less than 3% of violent crime. Property crime consists of burglary, larceny, and auto theft.

Though, I agree that these crimes are very different from violent crimes, where the motivation might be more like a problem with anger management or aggression control.

As to the "Results for murder are not robust if New York and the District of Columbia are included, but suggest a substantial elasticity as well", the author does later state that improvements in EMT, ambulances, and emergency room trauma care might account for some change in murder rate. Basically that crimes that 30 years ago were murders, are now assaults, just because the victim survived.

Originally Posted by casebro

There are a couple interesting tid bits in there Dinwar. I take it crime rates dropped in correlation with lead, but NOT in NYC and DC?

Regarding the New York and DC statistics differing from the rest, if I am understanding this right, the author is not talking about crime rates here, which do follow the same curve, but the lead per person calculations. Because NY and DC are so heavily populated (over ten times the mean), some results are skewed just by how they do the calculation, making it appear that New Yorkers breathed in way less lead than they actually did. This is why in some of the charts, the data is shown including NY and DC and excluding NY and DC.

Quote:

To account for (ii) the intensity of driving, we can multiply the gasoline lead content by total gallons of gasoline and divide by population to produce a measure of per-capita lead exposure: grams of lead per person. Unfortunately, this makes the assumption that lead is a purely private bad when in reality it has some public characteristics: a single gram of lead released on a crowded city street will certainly 14 affect more people than the same gram released on a deserted country road. For a high-population and high-density state like New York, the consequences of this mistaken assumption could be severe: not only does dividing by a large population produce an artificially low estimate of lead exposure, but failing to adjust for the high population density exacerbates this bias. These problems are likely to crop up for the highest population states (California and New York lead substantially) and the highest population density states (New York and the District of Columbia lead by far). As shown in Appendix Table 2, New York has the second-highest population and the single highest population density by a large margin.37 Because of this, New York ranks lowest on per-capita lead, even though it ranks only twelfth-lowest in grams per gallon and the population is densely packed and probably exposed to significant automobile exhaust. This issue would certainly encourage caution in the use of per-capita lead as a measure of lead exposure.

[meg]

Originally Posted by St.Michael

Is there any data avilable for Japan or other non-western countries showing a similar link?
This is interesting but could it be too good to be true, without considering other factors at play.

According to wikipedia, Japan didn't ban lead from gasoline until 1986, which, if the theory holds, would mean that Japan's crime rate should have started dropping in 2009.

Other countries in Asia didn't ban it until even later, some in the 2000s, so it would be too soon to tell, I guess.

http://en.wikipedia.org/wiki/Tetraethyllead

[Nessie]

After much study of poverty, inequality, age distribution and all sorts of reason for crime, along with all sorts of theories on how to reduce crime from 'Broken Windows' to access to abortions, we have the answer. Not only that we have been doing it for a good while now anyway.

Apparently taking the lead out of petrol and reduction in the use of lead for pipes is the cause of the spike in crime levels, particularly violent crime that saw crime rise from the 1950s to a high in the early 1970s, followed by a fall since then.

http://www.motherjones.com/environme...-link-gasoline

"In states where consumption of leaded gasoline declined slowly, crime declined slowly. Where it declined quickly, crime declined quickly."

This effect has been found to be the same in other countries as they phased out leaded petrol

".....lead data and crime data for Australia and found a close match. Ditto for Canada. And Great Britain and Finland and France and Italy and New Zealand and West Germany. Every time, the two curves fit each other astonishingly well."

This relates to the affect lead has on the brain such as aggression and IQ which has been known for sometime now and is correlated with the increasing rise of school exam success in the UK.

This follow on article deals with how the researchers are sure it is a causal and not correlated link.

http://www.motherjones.com/kevin-dru...onse-jim-manzi

The international correlation between introduction of lead and rise in violent crime, followed by removal of lead and subsequent drop is very compelling that the link is causal.

I am convinced.

[Irony]

First instinct is always "correlation does not equal causation" of course, but I've looked through the data and I'm fairly convinced as well. In this case it's not a one time correlation, but a rise and fall coinciding across a very large range of very different environments. Plus the mechanisms for how this could happen are not exactly mysterious and unspecified.

[casebro]

Mods, want to merge this wuth the other thread.in Sci & Med maybe?

[leftysergeant]

I would like to see statistics from China and Korea, where there is not only a high level of environmental lead, but mercury as well. Use of both lead and mercury and the incidental release of both elements are increasing in China. Korea is learning to take at least some environmental protection measures.

A lot of fancy dinner ware in both China and Korea, to some extent Japan, are red-laquered. The pigment is cinnebar, a primary ore of mercury.

[tyr_13]

Skeptoid blog has useful links at the bottom of the article.

As someone with ADHD, this interests me greatly.

[Nessie]

Apologies, I did not look far enough to find the other thread.

[Kuko 4000]

Steve Novella's take on Neuroblogica:

http://theness.com/neurologicablog/i...ead-and-crime/

Quote:

Conclusion

The connection between chronic lead exposure and neurological effects, including those that plausibly contribute to crime, is both plausible and reasonably supported by existing evidence. The magnitude of this effect is difficult to tease apart from the many variables that can potentially affect the crime rate. If we accept the 20% figure (crime that is lead related), which seems plausible, then this indicates a significant role for lead, but lead is certainly not the only important factor.

Also, because of the nature of this research there remains reasonable doubt about lead’s true role in the crime rate. This doubt, however, is not sufficient to argue that we should not pay attention to lead exposure or even take specific measures to limit it. The research is remarkably consistent in pointing to a real role for lead exposure. Multiple studies have also looked at the potential benefit of further reducing lead exposure (from soil and remaining lead pain, especially in window frames of old buildings). The research I can find all concludes that the benefit of lead reduction measures would be cost effective because of the potential benefits that would result.

[TubbaBlubba]

Originally Posted by Dinwar

I also am strongly biased against inferring causal relationships from mere correlation. I always want to see the actual mechanism. If you say "X causes 56% of Y" I want to see what the mechanism is, why 44% of the time the mechanism breaks down, etc.

That works fine in physics, but in social sciences, correlations are sometimes all we have to go by.

Luckily, there are statical methods (partial correlation and such) that can be used to infer causation from correlation.

[TubbaBlubba]

Originally Posted by Dinwar

At which point #3 becomes irrelevant. Why not simply discuss cognitive problems and learning disabilities? Why throw in crime rate at all? All it does, at that point, is confuse the issue.

Because crime rates are a useful indicator of certain cognitive difficulties. For one. They might be more interesting than cognitive difficulties.

Originally Posted by Dinwar

You have completely missed my point. Again, "crime" isn't a behavior IN ANY SENSE. Crime is violating laws. That's it. Show me how that is related to neurobiology and I'll shut up. Remember, we're not talking sociology here--the paper is postulating a link between low-level lead dosing and some BEHAVIOR, which makes this a neurobiology question.

No. It's postulating a link between lead exposure and social problems. The neurobiology is an attempt at an explanation.

1. "Look! Correlation!" (Sociological)
2. "Could this be the causative mechanism?" (Biological)

Quote:

The fact that sociologists use the statistics in no way demonstrates their validity. Most obviously, sociology addresses issues that are not the concern of neurobiology, such is income distribution, resource availability, and law.

What BIOLOGICAL ACTIVITY is being defined as "crime" here? Because if there's not one (or several) biological activities defined as "crime", the author is pulling a bait-and-switch: they're presenting their paper as a study in biology, but the paper is acutally about sociology. Not that sociology isn't a valid field of study, but once we accept that we're not talking biology we can abandon the pretense of looking for chemical causes (chemicals can impact behavior, but crime isn't a behavior, in any biological sense).

Also, the doorknob thing IS considered a crime. The American with Disabilities Act outlaws them (in practice, doorknobs cannot meet the standards required by teh AwDA, and therefore are illegal).

In the social sciences (and all other sciences, but it's most noticeable in social sciences), we sometimes "operationalize". This means that for the purpose of a scientific investigation, we attach a measurable definition to a more abstract thing, because you can't really study abstract things (like "crime" - what's the essence of a "crime"? Against the law? Might not be useful. Maybe we'll limit it to violent crime. But perhaps we shouldn't lump domestic violence with gang violence. And so on). Often you have to refine the operationalization as the research goes on.

Of course you know all of this already, and you're engaging in confusing rhetoric for no apparent good reason (Apparently because you don't like the link, for whatever reason). They are noting a chemical being correlated to a social problem. Obviously, the statistical correlation stuff will be sociological. The causative mechanism, if it can be identified (not a given) will be neurobiological. Different tools for different issues.

For example, if you want to know some detail about a fossil's chemical composition, you might employ chemistry, or an electron microscope. This is the exact same thing - if you want to know details of how something affects people on an individual level, you employ psychology and biology.

[Dinwar]

Originally Posted by TubbaBlubba

Of course you know all of this already, and you're engaging in confusing rhetoric for no apparent good reason...

Yeah, sure. It's not like I listed a few confounding factors with using crime rates, like the fact that "crime" has a shifting definition and the like. I disagree with the study, so OBVIOUSLY my motives are nepharious.

ETA:

Quote:

For example, if you want to know some detail about a fossil's chemical composition, you might employ chemistry, or an electron microscope.

This is actually a really good example. See, the chemistry of the fossil is impacted by numerous factors. The important thing with fossils is to determine that the chemistry of the critter is what you're actually seeing--rather than, say, an entirely different compound that replaced the original material. Similarly, we need to prove that at minimum the definition of "crime" remained constant throughout the period analyzed. Given the number of laws passed each year, I'd say that's something worth examining.

But of course, that can be dismisse because deep down, in my heart of hearts, I really do know the study is right, and my arguments are entirely assinine and not worthy of consideration. Or whatever crap you're going to post next.

[xtifr]

Ok, first of all, trying to guess what a paper said from a summary presented in the popular press is always a waste of everyone's time. The conclusion in Mother Jones is very likely quite different from what the paper actually said.

I haven't read the paper, but I did see another discussion on a different forum where a couple of people had read it. Most people (who had only seen the Mother Jones summary—which I also haven't read) were shouting "correlation does not imply causation", but the people who had read the paper were saying things like, "they're not claiming to have found causation, but their initial hypothesis, based on US state-by-state data, turned out to still fit perfectly when they began looking at world-wide data." Which is definitely an interesting result, if true, and definitely lends credence to the hypothesis that there is a causal connection in there somewhere.

Note that the paper, to the best of my knowledge does not say "lead causes crime", or any other such headline-grabbing nonsense. What it seems to say is that a strong correlation was found which suggested a causal link, and that further investigation strengthened, rather than weakened, that hypothesis.

To quote xkcd: "Correlation doesn't imply causation, but it does waggle its eyebrows suggestively and gesture furtively while mouthing 'look over there!'"

[casebro]

So what test could be done to bunches of convicts to cement the assertion? Remember, we want to know the prevalence of lead when they were developing as children. Bone samples? "Bones" on TV likes to use teeth to tell where a person grew.

[Jorghnassen]

Did they set up their DAGs properly?

[Dinwar]

Originally Posted by casebro

So what test could be done to bunches of convicts

Thank you for so elloquently expressing my problem with using crime rate: we don't actually have a way to MEASURE crime rate. We measure CONVICTION rate. Some people might add fudge factors to account for the difference. All of this introduces confounding factors, and those are things that are worth discussing.

To go back to TubbaBlubba's comparison between this and fossil chemistry: No one would consider it unreasonable to ask someone doing fossil chemistry research how they know that what they're looking at is actually the chemistry of the original critter. In fact, it's considered very bad form (as in, "This paper is worthless and should be rejected" bad form) to not include that data up front, before anyone asks it. So I don't see why asking how they know crime rate has anything to do with biology (the necessary assumption in order to tie lead exposure to crime rate--otherwise we're tying socioeconomic condition to crime rate) shouldn't be treated as a legitimate question.

Originally Posted by xtifr

To quote xkcd: "Correlation doesn't imply causation, but it does waggle its eyebrows suggestively and gesture furtively while mouthing 'look over there!'"

I'm not saying correlation doesn't imply causation, necessarily. I'm saying that there are factors that complicate interpretation of the data.

[Visual Purple]

Originally Posted by Dinwar

Thank you for so elloquently expressing my problem with using crime rate: we don't actually have a way to MEASURE crime rate. We measure CONVICTION rate.

This gets at a problem, but isn't quite right. No one in law enforcement or any field that uses crime data would use a conviction rate as a measure of the crime rate. For one thing, when crime goes up, arrest rates, particularly for lesser offenses often go down.

There are two sources that are used to my knowledge: police reports and victimization surveys. No one thinks police reports are much good as a measure, so that leaves victimization surveys, which have all the problems usually associated with trying to get a representative sample. However, because the surveys are not directly associated with law enforcement activity, they can (and, AFAIK do), use questions consistent across time, so that they measure something more or less constant even while legislatures mess around with criminal codes.

Also, the surveys and reports also separate out violent crimes, which really are, despite some changes in sexual mores possible affecting concepts of rape, a fairly consistent group of behaviors.

It's maybe worth noting that, although I am not particularly familiar with the statistics and academic literature, I've seen hundreds of rap sheets. Violent offenders are very often repeat offenders. It's not at all unusual to see someone who has committed multiple batteries and aggravated batteries before murdering (or being murdered). Property crimes--burglary, breaking and entering--are often also part of the mix. Or they just keep on with the lesser offenses. I typically don't learn much about the psychological profiles and academic records of those who commit offenses less serious than murder. With murderers, however, there's a very common pattern: special education and behavioral problems throughout the schools years, early drop out, and a long history of arrests. The arrests often involve assaults or batteries, but arrests related to substance abuse are typical too. If you told me that some pattern of neurological impairment was common in these offenders, I would not be surprised at all. (Probably some confirmation bias in these observations; I haven't been trying for a systematic study.)

(And off to sleep then work.)

[TubbaBlubba]

Originally Posted by Dinwar

Thank you for so elloquently expressing my problem with using crime rate: we don't actually have a way to MEASURE crime rate. We measure CONVICTION rate.

See, this is where "operationalization" comes in. A conviction rate would probably not be a very valid operationalization of crime rate, I'm guessing they're using multiple factors, including report rate, amount of money spent on law enforcement, surveys, and so on. The important thing is that they're constant over time and correlate well with other factors "crime rate" would be expected to impart.

Your objection is analogous to saying "We can't measure gravity, we can only measure the force it imparts on massive objects! How could any research on this POSSIBLY be valid?"

[TubbaBlubba]

Originally Posted by Dinwar

This is actually a really good example. See, the chemistry of the fossil is impacted by numerous factors. The important thing with fossils is to determine that the chemistry of the critter is what you're actually seeing--rather than, say, an entirely different compound that replaced the original material.

You're ALWAYS measuring something other than what you want to measure. If you use a scale to weigh something, you're not measuring mass - you're measuring some electromagnetic activity in the scale. LIDAR doesn't measure the speed of a vehicle, it measures the shift of a laser beam. Tests for hyperthyroidism don't actually measure "hyperthyroidism," they measure the amount of thyroid hormones in the blood.

And so on. The important thing is that the quantity measured correlates well with the stuff you're interested in seeing its effect on.

[Nessie]

It would appear reduction of lead has only affected violent crime levels, which would again match the affect lead has on the body.

The affect of reducing lead has been noticed in the UK, France, Australia, so different countries as well.

[casebro]

But shouldn't there also be a drop in some other measures of "insanity"? Commitment rate? Bi polar rate? Some other developmental disease state- maths scores rising?

Or is this concept proof that lead collects in the "criminal impulse control center" of the brain? Or maybe only in the "criminal impulse control center of the brains of minorities" ?

Oh wait, I think I have it- Lower lead exposure has improved development of the maternal instincts of inner city slum dwelling parents, and they are doing a better job of raising children. We can tell by the way the gang membership numbers have been dropping. Along with the incidence of piracy on the high seas.

eta, another scenario without any proposed mechanism: The lower crime rate is caused by Global Warming. Note how the global temps have peaked as the crime rate dropped? And, IT HAPPENED ALL OVER THE WHOLE WORLD.

Yeah sure.

[casebro]

Oh man, turn them all loose. They are not criminals, they are only sick. Lets start chelation treatments in the prison infirmaries. Call out the ADA...

[Soapy Sam]

So this isn't another gun control thread then? Oh pooh!