The EPA, Dust, And The Ecological Fallacy

In last night’s Republican debate, the candidates were asked if they could eliminate just one federal agency, which one would it be? Herman Cain chose the EPA: Herman Cain

The first — the first department, if I were forced to eliminate a department, I would start with the EPA and start all over. It’s out of control.
Now, I know that makes some people nervous, but the EPA has gone wild. The fact that they have a regulation that goes into effect January 1, 2012, to regulate dust says that they’ve gone too far.
So rather than try to fix it, eliminate all of the things that they have right now and then start rebuilding a responsible EPA.

Mother Jones jumped on the dust claim, calling it a “myth”:

Yes, the EPA is revisiting its dust standards—but those standards have been in place since 1987. In April, the EPA issued an evaluation of particulate matter pollution standards, because the report is a requirement under the Clean Air Act. And while the report suggested that dust standards should be tightened, the EPA has no plans to “regulate” dust any time soon.

In other words, the EPA already regulates dust and will “revisit” and “tighten” those regulations because the Clean Air Act mandates that they do. But saying the EPA wants to regulate dust is to speak a “myth.”

The EPA is concerned about “fine particulate matter“, or PM2.5, which are particles “less than 2.5 micrometers in diameter” and are “referred to as ‘fine’ particles and are believed to pose the greatest health risks. ”

One of the main constituents of health-risky PM2.5 is—are you ready?—dust. Dust? Dust is everywhere! My Goodness: What about the children! They’re on it: “One group at high risk is active children because they often spend a lot of time playing outdoors and their bodies are still developing.”

The key word above is “believed.” Why does the EPA “believe” dust poses “health risk”? The same way the EPA believes a lot of things: by relying—heavily—on the ecological fallacy, a statistical faux pas and leading generator of epidemiological over-certainty.

Last week we looked at a paper which announced the following distressing finding (the extensive criticism is here):

Cardiovascular disease (CVD) deaths, especially those from ischemic heart disease (IHD), are consistently and robustly associated with measures of fine particulate and traffic-related air pollution.

The sentence is plain and does not need much interpretation: fine particulate (PM2.5) and traffic-related air pollution is killing people.

Why did the authors make this claim? They must have designed some sort of experiment whereby people in various stages of health were gathered and the amount of dust that they inhaled over some period of time was measured. This might have involved CAT scans, in situ measurements of dust both in the air and in the lung, and so forth. Since one puff of dust probably isn’t deadly, these folks would have to be followed for a considerable time.

Just measuring dust would not be enough, of course. They’d have to measure, for every individual, their health and health history. Probably competent doctors, say, or nurses asked questions like, “Do you have heart disease? Does it run in your family?”, etc. Some people would be dead from heart attacks, so to ascertain whether the dust killed them, autopsies would have had to be performed.

Complicated business! But the health of children is involved, so no expense must have been spared. Yet did the authors actually do this?

No. Instead, the authors used a statistical model that first guessed where that person lived (at one time), then used another model to guess how much dust was in the air nearby where that person lived, then tried to see if these people died from heart attacks or suffered from other ailments. The model said, “Not significant!”, so they kept trying models (about a dozen of them) until they found one which gave them statistical “significance.”

The authors then invoked that miracle of certainty, the environmental fallacy. This fallacy (roughly) states that correlations are equivalent to causations. An (old) example might be to notice that in certain geographic areas ice creams sales are correlated with drownings. The EPA would use this correlation to say, “Ice cream causes drowning” and so seek to regulate it.

Never did the authors actually undertake any measurement that would allow them to make the claim CVD is “consistently and robustly associated with measures of fine particulate.” I stress that dust could cause CVD and other diseases, but the authors offered only tepid and error-ridden circumstantial evidence. Further, they offered no competitive theories for what might have caused the deaths they noted. Thus, the also fell pray to the “I can’t think of an alternate explanation so there isn’t one” fallacy.

Since this is only one of a legion of ecological-fallacy papers on which the EPA relies, Herman Cain was on to something when he suggested we scrap the EPA and start it again from scratch.


  1. Ray

    Mr. Cain is right. If the EPA doesn’t have any evidence, they just make stuff up. Remember, they declared DDT a class A carcinogen and banned it despite no evidence,

  2. Dr K.A. Rodgers

    Correlation and causation: All who drink water die.

  3. The lone downside to Mr. Cain’s suggestion is the resultant flooding of an otherwise limited employment market with thousands of incompetent scientists. But hey, let’s go for it anyway.

  4. Noblesse Oblige

    Cain was mild. We need to eliminate them all and start over. It is very clear that the central government is broken. The several States should exercise their Consititutional right to call a Convention to redesign their central government. Barring that, a regionalized US with common defense is the best bet.

  5. Rich

    Thus, the also fell pray to Well since they’re obviously superstitious then they are, equally obviously, wrong about everything else.

  6. Milton Hathaway

    Personally, I think the candidate that proposes sunsetting *ALL* federal regulations on a schedule that extends, say, over the next ten years, will have a clear winner of a campaign issue.

    If a federal regulation is still needed and beneficial, it should be able to stand on it’s own merits. The proponents/beneficiaries of a regulation should have to prove periodically that it’s still needed. In other words, guilty until proven innocent, unneeded until proven needed, and harmful until proven beneficial. Even if you love regulation, how do you argue with that logic?

    I’d only advocate this on the federal level, though. The states already have enough accountability, and the country gets 50 chances for them get it right. The federal government gets one chance to get it right, and of course never does, but the crap gets vested anyway.

    WD-40 is great stuff; WD-1 probably not so much.

  7. Roman

    How exactly do you propose the process should work? Do you expect the powerful bureaucrats to commit professional suicide? Or create another agency or a new tzar? This is the problem with democracy and we forgot that it supposed to be a republic. By definition it is impossible to overthrow a democracy (possible, but illegal), all you can expect is its growth.
    Can somebody model it and announce at what stage it will collapse? In the meantime we can enjoy scientific reports. Next: the assault on the the sector of fast food industry: federal standards for the binding of certain toppings to pizza crust.

  8. I like Noblesse Oblige’s suggestion. Make it a four year plan based on each agency’s budget impact for 2011, with all agencies divided into four unequally sized groups, each segment consisting of departments receiving an approx 25% slice of the federal funds in 2011, grouped low to high, – those receiving the least funds in the first group, etc; – effective July 1, 2013. Every federal agency will be sunsetted in the effective year, and will either reprove it’s worth and need to be reborn or reauthorized, or permanently closed down.

    Its unbiased and manageable, and gives existing staff time to prepare. What could be fairer? Imagine how streamlined some agencies could become? Imagine how many could be permanently closed? Imagine how busy Congress and the agencies will be and how little time they’d have left to consider stupid legislation and rule-making. Or at least I could dream.

  9. John A

    I am still waiting for the banning of bathtubs, showers, and other indoor plumbing – a start has been made, limiting the size of showerheads and the amount of water-per-flush, but surely there are still an alarming number of health considerations (accident falls, for instance) yet to be addressed.

  10. I agree with Noblesse… Cain was mild.

    As to the EPA, this just illustrates that a group of fairly smart people can construct a mathematical model to prove whatever fits their agenda. Garbage in = Garbage out.

  11. deadite

    This is an old problem, of which I have argued with many about. No, I’m not defending it. I do want to point out that there was a study (which I can’t find – it was a decade or more ago) that examined the populations of east and west germany prior to unification. The clean west germany had higher rates of asthma. The theory was that dust is a required ingrediant in developing our defenses, and that a lack of dust in the developing years leads to lungs that are unprepared asa the child gets older. I had many arguements with epa folk back then. They have been trying to get the PM2.5 since 1987 – the PM10 was all they could pull of at the time.

    I agree with Herman Cain. Send them all home, tell them they get two years severence, and start from scratch.

  12. I also agree with Cain. What he needs is a political ad to abolish the EPA with Queen’s hit song “And Another Bits the Dust” playing in the background.

Leave a Reply

Your email address will not be published. Required fields are marked *