The Link Between Lead and Crime

In the 1960s, a Caltech geochemist named Clair Patterson made the case that there had been worldwide contamination of living things by lead, due to the lead in gasoline. There were great increases in the amount of lead in fish and human skeletons, for example. More than anyone else he was responsible for the elimination of lead in gasoline. (By coincidence, this was just shown on the new Cosmos TV series.) A professor of pediatrics at the University of Pittsburgh named Herbert Needleman did some of the most important toxicology, linking lead exposure (presumably from paint) and IQ in children. Children with more lead in their teeth had lower IQ scores. The importance of this finding is shown by the fact he was accused of scientific misconduct.

When lead was eliminated from gasoline, blood levels of lead went down — and so did crime. The idea that childhood lead exposure causes crime many years later explains so many otherwise-hard-to-explain facts, especially worldwide declines in crime rates, that I conclude it’s true: lead exposure does cause criminality. Kevin Drum wrote a long article about this in Mother Jones a year ago and followed up his original article in many ways. A BBC radio show yesterday covered the topic.

This interests me for two reasons. One is simple. It shows the value of mercury in my teeth fillings made me stupider. There’s still lots of lead in the world — in old windowpanes, for example. And you are exposed to thousands of other modern chemicals (e.g., in cleaning products) whose effects on your brain are essentially unknown.

The other reason is complicated. It involves the context of this discovery. Mostly, the health research establishment has been unable to get anything right. Heart disease has been the #1 killer for decades; doctors still claim (and vast number of people, including New York Times health writers, believe them) that it is caused by cholesterol. Depression and bipolar disorder might be the single greatest cause of suffering nowadays — and psychiatrists are still claiming it is caused by “chemical imbalance” in the brain. (For my view of what causes depression, see this.) Beyond figuring out that lung cancer is caused by smoking, there has been almost no progress understanding what causes cancer. The “oncogene theory” of cancer turned out to be a dead end. There have been little bits of progress here and there but on the big issues, there has been nonsense decade after decade — and lack of realization that it is nonsense.

In contrast, taking lead out of gasoline was a big step forward in public health and pointing out the link to crime a big step forward in understanding crime. Rare examples of progress. What can I learn from that? I have stressed the importance of insider/outsiders — people close enough to understand but far enough away to have freedom. The lead/crime case supports that. Clair Patterson was a geochemist, not a toxicologist. Rick Nevin, the first person to argue that lead causes crime, was an economist, not a criminologist. Both of them had a good methodological understanding and used this to shed light on a different area than their original training. (Obviously I have used my background in experimental psychology, especially my methodological knowledge — how to experiment, how to measure brain function — to shed light on many health questions.) The lead/crime link also supports my view that the notion that “correlation does not equal causation” does more harm than good. The immediate response of many many people to the lead/crime evidence was exactly that — putting them on what turned out to be the wrong side. Whatever truth correlation does not equal causation might have is outweighed by the damage it does when it is used to ignore evidence. How smart do you have to be to realize “correlation does not equal causation” is stupid? To me don’t ignore evidence is the most important principle of science. But many university professors don’t agree with me.

I’m also impressed — in a good way — by Drum’s article. At least it exists. Anyone can read it and then look further, for example at original scientific articles. I wouldn’t say it was easy to write but it did not require expensive travel, extensive interviews, or months of research. It did require original thinking. In contrast, the New York Times and The New Yorker, which do allow expensive time-consuming journalism, haven’t published anything nearly as good in decades. The New York Times‘s idea of high-quality journalism seems to be a series about the high cost of health care while The New Yorker weighs in on the harm done by Dr. Mehmet Oz.





Thanks to James Keller.





24 Replies to “The Link Between Lead and Crime”

  1. Seth, what shows and Youtube channels do you (and other blog readers) currently use for providing a dose of “morning faces” ?

  2. I think you’re wrong about causation and correlation. The simple belief that correlation must imply causation does, I guess, far more damage than ….. well, than what? For many problems it’s perfectly reasonable to demand that an observed correlation be backed up by controlled experiments – whether on a population sample or on an individual. The problem arises where you can’t perform experiments – nobody is going to impose lead or cigarette smoke on a sample of people.

    There you can turn to Bradford Hill’s criteria – which seem to me superior to an uncritical supposition that correlation must imply causation.

    1. In my experience few people assume that correlations must imply causation. Whereas many people use “correlation does not equal causation” to ignore evidence.

  3. the complete state should read: “Correlation does not imply causation, but it does demand explanation.”

  4. > The problem arises where you can’t perform experiments – nobody is going to impose lead or cigarette smoke on a sample of people.

    Dearieme, yes yes, everyone knows you can’t experiment on people – *that* way. There’s still ways you can try to experimentally confirm it: for example, instead of focusing on *adding* lead, focus on *removing* it; you can go find an area with lots of lead in its building, pay for random buildings to be renovated and de-leaded, and monitor outcomes of their inhabitants. This will require a lot more money and data than if you could add big doses of lead (since generally there’s not that much lead you can remove so the effect will be small), but it’s doable if there’s will.

    Unfortunately, there isn’t. Most people don’t appreciate the value of this sort of research. (How much would an experimental confirmation or debunking of the effect of lead on crime be worth over the next century? Billions, at a minimum. Is anyone running these sorts of experiments? Nope.)

  5. People who think the “correlation does not equal causation” principle does more harm than good, would certainly DEMAND that they NOT be tried for a capital crime (i.e., punishable by death) in a court of law, based solely on correlation. Or I would hope not.

    And real lives are also at stake here.

    There would be a demand for EVIDENCE, wouldn’t there? Just because there is a far stronger correlation between men and murder than women and murder, a good criminal investigator would never automatically discount the idea that a woman may have committed a particular murder, and not a man. He or she would still look for EVIDENCE, right? Just imagine how many men would unjustly lose their lives if it were otherwise.

    It’s the basis for the Hippocratic rule of: First, do no harm.

    The reason so many scientists and researchers like to use correlations rather than actual scientific evidence is this: $$$

    And it’s easy to find a correlation. But actual EVIDENCE? Not so much.

  6. In grammer school I was taught that the ancient Romans used lead pipes in their cities for plumbing, which poisoned them. Later I discovered that most if not all major cities in the United States use lead pipes throughout the cities to distribute water to homes. The city water, of course, is also filled with poison: arsenic. chlorine. clorates and flouride (flouride is rat poison). Many of the weights that bodybuilders and weightlifters use are filled with lead, with no lead warning label whatsoever on them.

    Either city planners never went to school, refuse to learn from history, or don’t care; or there is a gigantic conspiracy to keep people sick to keep them physically and mentally weak and submissive and trapped in the modern health care system, which is actually a Death Cult. It is amazing that any of us are still alive, considering all of the poisons that exist in our air, water, food and “medicines”.

    Good article!

  7. Seth,

    I hope this is not a hijack but I had a few questions about the practicalities of morning faces. Apologies if you have covered these questions elsewhere, but I’ve read through your (extremely helpful!) summary page and couldn’t find answers.

    1) When you say morning, have you or others found that earlier is better? I get up around 7 am but as a practical matter, it would be easiest to do this while at my desk, around 9:30 or 10. Of course I will endeavor to self-experiment and find out, but would be curious to know if you have any thoughts on this question.

    2) Do you find that you need to be staring directly at the faces for this to work? I ask because as a practical matter it would be easiest if I could have this running on my computer screen while I’m doing other things.

    1. Is earlier better? There is an ideal time, where moving the faces earlier or later produces weaker results. The only way I know to figure out the ideal time is trial and error.

      Do you have to be staring directly at the faces for this to work? Yes. My guess is that the sensitive period is roughly 2 hours long. During this period you need at least 30 minutes (ideally 50 minutes or more) to get a big useful effect. So you need look directly at the faces for at least 30 minutes during the sensitive period. While looking at the faces you can listen to whatever you want — music, podcast, book, etc.

  8. Thanks for the pointer to bloggingheads.

    Did you ever do any experiments with looking at still images of peoples faces versus videos to see if there was any impact ?

  9. The only problem with removing lead is that lead has been replaced with other additives that mimic the behavior of lead in controlled gasoline combustion.

    As a chemist friend of mine said, okay remove lead, remove that as one source of stupidity and replace with partial pyrolysed benzene and other aromatic ring structures in early morning start-ups and by exhaust catalysts that are sub optimal.

  10. 30 minutes sounds like a long time to stare at faces. How about getting a dog? Your unconscious can’t tell the difference between human and dog faces, right? The assumption is already that it can’t tell the difference between virtual and real faces.

    1. I can’t imagine staring a dog straight in the face for 30 minutes. 30 minutes isn’t a long time for a conversation.

  11. But, gwern, it’s so easy for your sort of experiment to miss the point. For example, people gathered the statistics for one of the big Australian lead-mining towns. (Broken Hill, I imagine.) People there were as healthy as you like – no sign of lead cutting a swathe through them. But, it seems likely, that if lead in gasoline was the problem then you need to study not the rather large particles (“dust”) emitted by ore mining, but the tiny particles emitted from internal combustion engines.

    dr j: “As a chemist friend of mine said, okay remove lead, remove that as one source of stupidity and replace with partial pyrolysed benzene and other aromatic ring structures in early morning start-ups and by exhaust catalysts that are sub optimal.” That may be why there was then a war against aromatic compounds in gasoline. Benzene, once used to wash grease off your hands in the chem lab, was discovered to be far more dangerous than had been understood.

  12. “The idea that childhood lead exposure causes crime many years later explains so many otherwise-hard-to-explain facts, especially worldwide declines in crime rates, that I conclude it’s true: lead exposure does cause criminality.”

    But that might extend to having exposure to lead being used to explain every otherwise inexplicable rise or fall over time. Does decline in lead explain rises in divorce rates, bastardy, and so on?

    Does the lead argument satisfy Bradford Hill’s criteria?

  13. For pure health reasons, there is plenty of cause to “get the lead out.” We need to do more and faster.

    But that is different from pinning the blame for violence on lead.

    Here is an article which takes a skeptical look at the lead-violence link. The issue is not so obvious. The article gives academic references.

    Here is another idea. Modern chemicals reduce testosterone in males, and that reduces aggression.

    And another theory. It is well known that heat, particularly humid heat, put a burden on the body, which must keep the brain cool. If it is too hot, people become “hot heads.” An overheated brain will far more easily commit acts of violence. It is why violence is higher in summer. The widespead adoption of air conditioning and new dress standards which allow cooler clothing may have reduced violence.

    Could not greater incarceration contribute to lower social violence (outside of prisons, that is)?

    New York has had an exceptionally large drop in violent crime. Is it not PC to say that aggressive policing made a difference?

Comments are closed.