Lead and violent crime — why a good hypothesis isn't proof

We know that lead exposure can be dangerous. We know that it can cause brain damage. But what levels are dangerous. How does that damage express itself? And how do you separate the effects of lead poisoning from a whole host of other potentially dangerous, damaging factors? Last week, Mother Jones had a well-done article about research that is drawing connections between leaded gasoline and the crime wave of the mid 20th century. That's a hypothesis. It's a hypothesis with a lot of correlational evidence. But it's not proof. I recommend reading public health researcher Scott Firestone's excellent article that delves into the details of the studies from the Mother Jones story. It's a great look at the lines between public health as a science and public health as activism and it helps shine some light on why seemingly airtight cases aren't always immediately acted upon.


  1. There is a legitimate (also disputed) case made that a very strong contributing factor to the fall of the Roman Empire was an explosive rise in violent criminality among its citizens, which can also be correlated to their use of lead in many aspects of daily life, including pipes to bring water into their cities. Those who dispute this say that lead was already known to be dangerous back then, but that doesn’t mean people didn’t use it anyway. If we think it’s hard to educate people and enforce environmental regulations today, just imagine the crazy shit people used to do back then.

    1.  Why not the *rise* of the Roman Republic?

      A major problem with your theory is that they used lead acatate as a sweetener throughout their history.

      If anything, aristocrats in the city may have had more access to lead acetate than anyone else, and the Julio-Claudian dynasty, which was composed of aristocrats from the city had some people with serious problems. But the later emperors were often soldiers from the frontier and they tended to have different problems.

      1. Lead was widely used at all levels of society, not just for sweetening the wine of ruling class aristocrats. It was used as a sweetener and preservative for all kinds of foods. It was used as a cosmetic, it was used as medecine. It was used in plumbing. It was the preferred metal for certain types of cookware. If not used in pure form, it was in metal alloys, and it was also used as lining for pots made of other materials. It was used everywhere. And there were no environmental agencies monitoring how its industrial extraction and processing (smoke, water runoff, waste dumping) was conducted.

        Sure, as a versatile material that was integral to so many technological and cultural activities, lead certainly did play an important part in the rise of the Roman Empire. But by virtue of its ubiquity, and the way it accumulates in the body and in the environment, it would be silly to dismiss it as a clear factor in bringing Roman society down.

        1. My point is that it must have had nasty effects throughout society, and perhaps with some bias towards the most privileged who could afford the most abuse, but that these nasty effects began far too early to explain the collapse of the Roman Empire, unless you can explain why the lead poisoning was more important in the late Roman Empire than in the late Roman Republic.

          1. “unless you can explain why the lead poisoning was more important in the late Roman Empire than in the late Roman Republic.”

            since you’re asking, my theory is that it is because it takes time for the effect to build up, it became more of a factor as the environmental exposure increased along with the growth of the society which made such ubiquitous use of lead in all aspects of daily life.

          2. It doesn’t take seven or more centuries for the effect to build up, the example above took less than fifty years to build up. And ancient agricultural societies tended towards stasis not growth. It is likely that agricultural productivity and population were declining due to soil erosion.

            Edit: revised based on 3rd century B.C.E. to 5th century C.E.

          3. did I miss something? did the united states fall into barbarianism over the course of those 50 years? or was there simply a measurable increase in crime?

          4. With tetraethyl lead pollution, you have a gradual increase in tetraethyl lead pollution, and a gradual increase in violence afterwards, and then a gradual decline in tetraethyl lead pollution, and a gradual decline in violence afterwards.

            With lead acetate consumption, you have the initial introduction of lead acetate sweeteners, and then a relatively stable level of lead acetate consumption for several centuries, and then a sudden institutional collapse, and I think a continuation of that relatively stable level of lead acetate consumption.

            Can you explain why the 25th generation of lead acetate users would be affected any differently from the 20th generation of lead acetate users? Lead accumulates over a lifetime, but I doubt it accumulates over several lifetimes. Lead poisoning explains delayed effects 20 years later but I doubt it explains delayed effects 200 or 400 or 800 years later.

            P.S. I think Sapa was introduced by the 3rd century B.C.E. but if someone is more familiar with the subject, can they clarify?

  2. All too often correlation is used to claim causation by a Political Class who finds such claims beneficial to their and their backers’ ulterior interests in the new policy.

    1. Nobody in their right mind would bring TEL back… I don’t know who might conceivably benefit from this logical mistake now, unless a large class action lawsuit was being thought of.  The only remaining TEL users – older aircraft where the valve seats need it – are working slowly to phase it out or refresh engines.  Already.

  3. I hope no-one thought the MJ article was proof! They even said in the article that they need more study.

    But I have to say, I’d love to see the lead hypothesis borne out. It’s just too good.

    1. Yeah. I thought the whole thesis of the article was “we should probably be studying this more.”

      What’s a science journalist to do? Underselling your own case doesn’t exactly make for riveting content, but everybody goes into full on debunking mode if you are too confident.

  4. Has anyone said that this hypothesis is proof?  The article is laying out a hypothesis and showing why it’s an avenue which is worthy of being examined.  This is the basic first step in science.  So – you’re suggesting that Mother Jones should not publish a hypothesis because it’s not accepted science yet?  That’s a strange bar to set with the state of most journalism today…

    1. No. I’m not suggesting that at all. I’m suggesting that, as a consumer of media, it’s helpful to have multiple perspectives on complicated questions like this. Far too often, I see myself and my friends read something that’s really persuasive and well-done, like the Mother Jones story was, and when we talk about it to other people though what comes out is “X caused X”. 

      I think that having multiple perspectives both helps to avoid that cocktail party simplification that happens so easily and helps us to understand what’s going on in the background of things like this. Sometimes, there really are corporate conspiracies taht keep dangers like lead from being acted on. Sometimes there’s not a conspiracy and there’s simply not enough evidence to convince action. Sometimes it’s a combination of both. 

      1. One of the things I like most about your blogging on boing boing is how consistently you revisit stories you’ve already mentioned with competing perspectives or hypotheses.  The world is a confusing mess and I appreciate that you take this fact seriously.

  5. Clearly, this is a dispute that only unethical experimentation on human infants can resolve!

  6. Interesting comment from your linked site Maggie. It boils down to how to determine the burden of proof. Should studies start from a blameless perspective, a proof that no harm was done, or somewhere in the middle?

  7. Kevin Drum, didn’t say that there was an interesting correlation. He went on with all the “blindingly obvious” rhetoric. He went from scientific enquiry into rhetorical scientism. In essence, as in many claims of these sorts, he gave lip service to the idea that correlation does not necessarily equal causation, but then went to imply just that. 

    Causation studies is the basis for the scientific method, precisely because of how unreliable base correlation is. 

    Anecdote or intuition can lead to a correlational study which can then lead to a causational experiment. But until we go through the entire process, it ain’t science. 

    Sometimes we don’t have the time,or we lack a methodology, or we lack the tools to do science. In those cases we will have to do our best without science, we will proceed the best we can. But in those cases, it is best to be precise, that we studied the problem yes, but not scientifically.

    1. Well, to be fair I think Drum did do a good job of arguing there was evidence beyond the correlation that made this a pretty good hypothesis and hedged himself enough to be able to say he never claimed it was more than a hypothesis.  But you’re absolutely right that he was employing rhetoric to drive the reader to the “lead is at fault” conclusion.

      I thought it was OK since the lead hypothesis does seem to be neglected in the attempt to explain patterns of criminal behavior in the late 20th century.  Scientists also tend to root for particular hypotheses and good science happens when different scientists back different hypotheses.  So strongly arguing for a particular hypothesis should not be taken as being inimical to science — in some sense it seems necessary for it.

  8. The original Mother Jones article was an excellent read and gave me a new perspective on the horrors of lead.  It did strike me as a tad over confident with statements like these:

    “Put this all together and the benefits of lead cleanup could be in the neighborhood of $200 billion per year. In other words, an annual investment of $20 billion for 20 years could produce returns of 10-to-1 every single year for decades to come. Those are returns that Wall Street hedge funds can only dream of.”

    As Maggie says, more proof is needed and I don’t think we are ready to spend $20 billion a year on this.  We have many more certain things that we need to attend to first.

  9. It seems more likely to me that the crime wave was a cultural ripple effect of World War II.  But disproving a hypothesis like that would be nigh well impossible.  So it will remain an article of faith for me.

Comments are closed.