Hacker News Reader

Comment by coffeecat | Hacker News Reader

This topic is one of my biggest pet peeves. The hypothesis that sub-clinical lead exposure _causes_ IQ loss is, in my opinion, much weaker than most people realize. Observational studies (which are more or less all we have) tell us that Pb exposure correlates negatively with IQ, and that the correlation is moderated by a multitude of variables that are subtle and difficult to model. These studies do not tell us in which direction the arrow of causality points - they merely tell us that not all causative pathways are accounted for by the study's model. The idea that causation points from lead to IQ loss has some issues - e.g., that developmentally delayed two-year-olds ingest more dust/dirt, which increases their lead exposure.

View on HN

I feel like you maybe should shop around for a new pet peeve. Wasn't the data on this pretty clear due to various countries, states, cities banning lead at different times? My understanding is that you can see a corresponding drop in lead levels in the blood, rise in IQ, and reduction in crime 20 years later.

https://en.wikipedia.org/wiki/Lead–crime_hypothesis#Correlat...

Tying lead levels to crime levels seems super sketchy to me when there are plenty of other factors that get studied and seem to have much better correlation and even causative effects. Hell, violent crime has started creeping back up in recent years, but to my knowledge the lead levels are not.

Violent crime "creeping up" is entirely dependant on whose reports and which methodology relative when it changed .. see (for example):

https://www.themarshallproject.org/2023/11/03/violent-crime-...

Whichever way that data is cut, though, pre-2000 and in particular mid 70s "peak lead" crime levels are far and away worse than those of the past two decades.

That's fine, but I don't think there any causation in that correlation. I've never heard a compelling argument to explain it when there are tons of other compelling factors (organized crime, economics, opportunity, surveillance technology, testosterone levels, even birth control and abortion).

Edit: why disagree? I can claim that the size of the standard Hersey bar had a positive correlation over that time too. That doesn't mean there's a convincing causative mechanism. Most of the studies linking lead to crime rely on poor data practices, such as testing blood levels of convicted criminals. There are very easy explainations to show bias here, such as smarter criminals being less likely to be convinced and criminals with higher blood levels tending to have lower IQs. Many exclude certain areas like DC or NYC in their analysis or the trend disappeared. Many used state level data points which are not granular enough to investigate causative effect (eg median income at the state level is not useful. You need to investigate it at the individual level. Things like localized cost of living and standard of living would also be more useful numbers). The designs and data are sloppy as hell in my opinion.

Yeah that's not how it works though... nobody is saying lead causes violent crime. Nobody is claiming crime can't go back up without lead. Of course there are better causative factors.

It's just that meta-analysis over many studies seems to show that it is a contributor to crime. When it is removed from the environment, it then shows up in the crime data.

"nobody is saying lead causes violent crime...

many studies seems to show that it is a contributor to crime."

This seems contradictory. How can it not cause violent crime yet be a causative factor?

I can see how it might be correlated. But I've never heard a compelling argument for how it would be causative.on any significant level.

Sorry, When I said "nobody is saying lead causes violent crime" I meant "nobody is saying lead is the sole cause of violent crime."

Please do share if you're aware of any sources associating population-wide lead level decreases to population-wide IQ increases. While we can't be certain about blood lead levels before the mid-20th century, my impression is that the golden age of the Flynn effect (mid-20th century through the mid-70s) was concomitant with explosive growth in the use of leaded gasoline, which should be far and away the primary driver of elevated lead levels, and that cognitive scores have been relatively flat from the mid-70s onward as lead levels have decreased precipitously.

Edit: It's worth noting that the study described by the OP does not demonstrate an effect on IQ. They estimate population-wide lead exposure. The lost IQ point calculation is based on correlation coefficients obtained through observational studies.

There is a basis in fact and theory that the Roman Empire fell, in part, due to Lead Acetate (AKA sugar of lead) being used as a sweetener. As far as I know, the bad effects of lead are well established, and the use of tetra ethylated lead as anti knock additive in car and aviation has endured in the face of solid science = lead is toxic. The oil and auto business bribed regulators to ignore the science, much like the tobacco bribe industry. I use the term 'bribe' in the true sense, https://en.wiktionary.org/wiki/bribe, versus the political sense where 'lobby' is preferred. https://en.wikipedia.org/wiki/Lobbying There is also the Tammany Hall origin - I will meet you in the lobby with a bag of $$, which some say has been done in Congress in those bad old days.

Models are hard and causation near impossible, but feel pretty safe being on the "arrow of causality pointing from lead to low IQ" rather than the hilarious inverse...

> the hilarious inverse

Made me think of this:

1. Make up some scientific "evidence" that exposure to some common, harmless substance is actually harmful.

2. People believe it. Smart people make an effort to avoid it. Dumb people don't care.

3. Measure it 30 years later, and now there's a correlation between higher IQ and lower exposure to the substance. Higher IQ causes lower exposure to the substance.

I'm not saying this is the case with Pb, it's just a funny idea.

There are longitudinal studies that track people from childhood before the dangers of lead were widely known. There are also comparative studies where populations are studied with respect to occupational exposure.

  Smart people make an effort to avoid it, dumb people don't care.

With lead how would a smart person do this? Move from the US to a country that banned leaded gasoline earlier?

Is your sense that the number of smart people leaving the US each year is greater than the number of smart people moving to the US each year?

Not parent, but:

> I'm not saying this is the case with Pb, it's just a funny idea.

https://en.wikipedia.org/wiki/Human_capital_flight

Not nearly as hilarious as the idea that lead is more harmful at levels <5 mcg/dL than it is in the 10-20 mcg/dL range. But that's the only way to explain the available data while attributing the observed correlation to Pb neurotoxicity. And indeed, it's the mainstream view among lead-IQ researchers.

Back when lead levels were mostly in the 10-20 mcg/dL range, the observed correlations with IQ were typically around 3-5 IQ points per standard deviation of lead exposure. Now that lead levels are an order of magnitude or two lower, the observed correlation is still... 3-5 IQ points per standard deviation of exposure. Blood levels of other environmental toxins have also been correlated with IQ. These studies typically come in around... wait for it... 3-5 IQ points per standard deviation of exposure.

The common-sense explanation, in my opinion, is that developmentally delayed children eat more dust and dirt.

The implication is not that low IQ causes lead exposure, but that low IQ and lead exposure are both caused by a third factor, presumably poverty.

Poverty is easy to control for. Controlling for it brings the Pb-IQ effect size down considerably, but not down to zero.

My view is that poverty, low IQ, and lead exposure are all caused by intractably complex webs of genes, behaviors, beliefs, and parenting styles. This occurs in such a way that high IQ, high socioeconomic status, and low lead exposure are all strongly correlated, but not perfectly correlated.

As a spherical-cow sort of model, imagine that there's a "I don't give a fuck" (IDGAF) gene. People with the IDGAF gene have higher levels of lead exposure (because their houses aren't as clean), lower IQs (because they're not as motivated to think hard about the IQ test problems), and lower socioeconomic status (because they're not as motivated to get high-paying jobs) relative to those without the gene. But some IDGAF-positive people will get lucky and end up with high-paying jobs. Some IDGAF-negative people will be impoverished due to bad luck, etc. In other words, socioeconomic status is a noisy indicator of IDGAF. The key point is that IDGAF-positive people who have good jobs still probably don't keep their houses as clean as IDGAF-negative people. So when you control for poverty, you've killed off most of the IDGAF-mediated correlation between lead and IQ, but not all of it.

(as a reminder, this was just a spherical-cow model for illustrative purposes... reality is intractably more complex!)

This isn't any old epidemiological data mining exercise. We know that lead directly causes IQ loss in larger amounts. It should take a lot less evidence to convince us that correlations pertaining to smaller exposures are therefore causal.

I've got two things to say about that:

1) Multiple researchers have obtained U-shaped lead-IQ curves which regress toward the mean at extreme values. This, in my opinion, strongly implies that the lead-IQ correlation only occurs in the "ordinary exposure through dirt and dust" regime.

2) The effect sizes reported back in the 70s, when lead levels were an order of magnitude or two higher than they are now, were significantly lower than the effect sizes reported in contemporary research. In both cases, however, they can be broadly summarized as "3-5 IQ points per standard deviation of exposure".

That's the midwit's causation vs correlation meme which is usually at play when an HNer uses both those words in a comment.

Observational research is how we know that smoking cigarettes causes health issues. We don't have 30 year RCTs on smokers. And there is no such thing as causation vs correlation. All we have are causal inferences draws from correlation.

Finally, when you suggest there are confounders, that's a causal claim. What confounders do you have in mind here, what standard of evidence do you need to accept them as confounders, and do they supersede the evidence we have for lead and IQ?

"And there is no such thing as causation vs correlation. All we have are causal inferences draws from correlation."

A lot of it depends on how well the controls were done and what follow-up research was done. Thongs like identifying and confirming mechanisms of action are a huge difference between some basic correlation and claimed causation. In the case of exposures, you can work backwards - reduce blood levels and perform tests to see if functions return. That's a lot easier than identifying people before they meet the criteria. Although that's still possible, especially when dealing with certain vocational scenarios where tests can be performed before possible exposures.

“Reduce blood levels and see if functions return” so like reduce smoking to see if lung cancer go into remission? Why the assumption health effects are reversible?

You can certainly reduce smoking to see if lung capacity improves. I don't think it's reasonable to reduce smoking to see if that will cause an existing tumor to go into remission, for the same reason it's not reasonable to see if installing a blade guard on a saw will make a amputated finger reattach or regrow. The risk has manifested, and removing the risk factor is too late.

You'd need a large study to determine if stopping smoking reduces future risk of tumors; it's not something you can determine in a single person.

"You'd need a large study to determine if stopping smoking reduces future risk of tumors; it's not something you can determine in a single person."

And they have done some studies that seem to prove the risk does decrease after cessation and continues to decrease over time.

Not all health affects are reversible. Some reversals can be quite evident for acute symptoms, like arsenic. IQ reduction from lead appears to be permenant. Using your example of lung cancer... no, the cancer would not go into remission if established. However, there are studies showing lung cancer risk go down for smokers after they stop smoking, and generally the risk continues to decrease the more time has passed since they last smoked.

"Correlation does not imply causation" is a simple idea, but it truly cannot be emphasized enough. Establishing a correlation certainly does help to strengthen an already-compelling causal mechanism theory. Pulling high doses of carcinogen-filled particles into your lungs causes lung cancer? Sounds about right. Reduced IQ from exposure to a neurotoxin at levels significantly lower than that which produces clinical symptoms? Maybe, but I'd say the burden of evidence is higher on that one.

> Finally, when you suggest there are confounders, that's a causal claim. What confounders do you have in mind here, what standard of evidence do you need to accept them as confounders, and do they supersede the evidence we have for lead and IQ?

Dirt and dust ingestion, for which lead is a noisy metric. It's no accident that the correlation between lead and IQ is strongest at age 2. That's the age by which the smarter children have figured out that dust bunnies don't taste very good.

You have a point in that people should continue to study this stuff. But, nobody is halting research here?

Do you have thoughts on why such a strong correlation would appear other than some causal influence?

To be sure, there could be other factors at play. But nobody, I mean nobody, is suggesting that lead is safe. We know it messes you up. There is evidence that it really messes you up.

There's been an astonishing amount of research done on the lead-IQ relationship over 50+ years. It's no exaggeration to say that this is one of the most thoroughly investigated public health topics. And one of the most interesting.

My belief is that lead is a noisy metric for dirt/dust ingestion, which correlates with IQ for a variety of reasons. Homes of poor families tend to be dirtier, and wealthier families tend to live in newer suburban housing farther from curbs for example, but that sort of thing is pretty easy to control for. The big thing that can't be controlled in a straightforward way, is developmentally delayed children eating more dust/dirt. To a much lesser extent, this also correlates negatively with IQ in adults due to pica, disregard for cleanliness, etc correlating negatively with IQ. But the lead-IQ correlation is strongest at age 2, and that's no coincidence.

Observable symptoms of lead poisoning occur at blood levels above ~60 mcg/dL, which is 3-4 times higher than the levels studied in 1970s-era lead IQ research, and about 2 orders of magnitude higher than the levels studied in contemporary lead-IQ research. The latter body of research reports significantly larger effect sizes than the former.

Would really like to hear your theory about how IQ loss causes lead poisoning.

In well-controlled studies, primarily via developmentally delayed children ingesting more dust/dirt. It's also worth pointing out that actual lead poisoning occurs at levels above about 60 mcg/dL, is very rare nowadays, and that the lead-IQ curve actually regresses toward the mean once you approach that sort of extreme lead level.

I'm sure the scientists considered "correlation does not equal causation" before publishing.

You'd think that, but...

I'll follow this with a more substantive comment. Science is a machine. To get papers published, pass your dissertation, win grants, and get tenure, there's an easy road and a hard road. The easy road is: "in conclusion, we found that lead correlates negatively with IQ, in support of the prevailing view on Pb neurotoxicity at low lead levels." The hard road is: "in conclusion, we found that lead correlates negatively with IQ, but correlation doesn't imply causation and the consensus view has some serious issues."

The article that the OP was written about is case in point. One calculation they do in their paper is the sum of all IQ points lost due to Pb exposure, which comes out to some absurdly large number. Sums of IQ points don't mean anything. It's just a sensational marketing gimmick. And it worked! It got them a national mainstream news article. Reporting their findings on lead exposure without the sensational IQ point calculations and commentary probably would not have gotten them a national news article.

I know that the people still trumpeting about the relationship between college degrees and income definitely don't. Ahem, Destiny, ahem, ahem

Reply on news.ycombinator.com