Causes in search of problems
Spurious associations represent the most glaring of all reasoning mistakes, and they are ubiquitous
Seen elsewhere online: a table with three columns, the first one labelled “Planets”, containing, as one would expect, “Mercury”, “Venus”, “Earth”, “Mars”, etc., all the way down (or up) to Neptune. On the second one, the label is a question: “Has any human ever died there?”. Over the third, another question: “Are there birds?” The only planet that gives a “yes” to both is, of course, our own good old Earth. Hence, birds are the cause of human mortality.
It's a joke, of course, but it’s also kind of profound. Imagine the headline: “Human mortality strongly associated with presence of birds in the biosphere”. It could be in the health and well-being section of The New York Times.
Of course, the truth is that human mortality is inextricably linked to the existence of humans, and humans, to exist in great numbers and without special equipment, require the kind of biosphere we have here on Earth. A biosphere that includes, among many other things, birds. The same headline could have been construed with lizards, flowers, beetles, fish, or trees.
Theoretically, everybody knows, or is capable of intuiting, that things that go together don't necessarily have a cause-and-effect link.
They may both be effects of the same underlying cause; they may simply be independent phenomena that happen to move in the same direction over time (such as the growth of a newborn child and that of the neighbor's puppy). They may even have no connection at all, the association being a coincidence. In some circles, “Correlation does not imply causation” is as tired and old a cliché as “a tired and old cliché”. You say it, and the people “in the know” will just nod and sigh.
Besides, there are well-known rules about moving, in an intellectually responsible (not to say minimally honest) way, from noting a coincidence to affirming a causal implication. Ideally, you need a way to isolate or, at the very least, provide a proper account of all extraneous factors that could be relevant to the issue. If doing so is impossible or impractical, you will need, at least, to present strong indicators of plausibility.
One may expect, for instance, a time sequence (the purported cause always precedes the purported effect), consistency (the same association is observed in different scenarios and conditions), and empirical viability (there is a logical or mechanistic basis supporting and explaining the causal claim).
But, despite all this, you can find shoddy speak of the form “I think I saw A and B together, so obviously B is the merit/fault of A” being used as if they were bona fide arguments everywhere. Pundits and politicians live and die by it. Economists, social scientists, and conspiracy theorists build massive castles on top of such flimsy clouds. And, of course, there are the speculative epidemiologists.
Speculative epidemiology is what gives birth to the phenomenon I chose to name this column after — the cause in search of a problem. Unlike the classic correlation-causation fallacy, it doesn't jump from an apparent puzzling coincidence, already evident in the real world, to an invalid inference.
Instead, it begins with something someone doesn't like — it can be people from a certain race, religion, or gender, or a chemical, or a technology, or an art form — and then goes out fishing for something, anything, that may be construed as a suggestive coincidence. You start with the cause, and then shop around for the purported effect. The only criterion: the effect has to be bad.
In the social sciences, examples would be the race-IQ baloney and the several youth-violence moral panics (comics, videogames, heavy metal music…). But the most fertile ground for speculative epidemiology is, of course, human health. More specifically, scary (and, hence, purportedly newsworthy) allegations about things that are going to ruin human health.
A few decades ago, we had cellphones and brain cancer. For a while, contraceptives and depression. Then, the herbicide glyphosate, GMOs and, really, everything that’s evil under the sun. Then, “endocrine disruptors” and something very bad but still very undefined, the same category as micro- and nanoplastics. With the advent of MAHA, the association between food additives and cancer is now in vogue again. Fluoride in the water and vaccines have also been targeted by efforts to “promote” them to universal villainy.
The list of things that are bad for you, even if nobody knows exactly why and how (or, conversely, for any reason and in any way imaginable), grows every day, published in scientific journals and amplified by the health and wellness press—so much so that the purveyors of “detox” quackery, the greatest beneficiaries of all this activity, struggle to keep apace.
A combination of the naturalistic fallacy (natural is always good, artificial is always evil) and a diffuse but well-ingrained antipathy towards powerful agents (big corporations or the state, depending on your ideological proclivities) tends to guide the choice of candidates for the list. Truth to be told, given the dismal track record of corporations in dealing ethically with the health hazards their profit-guzzling creates (the cases of leaded gasoline and tobacco being the most notorious, and also sugary sodas), the bad will is not altogether misplaced. But it can be, and often is, misapplied.
The problem with problem-hunting — the frenzied activity of trying to fit a cause to some dire effect, any dire effect — is that sooner or later it will succeed by pure dumb luck. If you look hard enough for coincidences that appear meaningful, you’ll find them. However, there will be a big chance that the meaningfulness will be illusory.
Imagine a worldwide championship of coin tossing, with a little over 1 million participants. At the first round, everybody flips a coin. Those who obtain heads stay, those with tails are eliminated. The winners then flip again: heads stay, tails out. And so the next round, and the next. The champion — the last coin-flipper standing — will have accumulated a sequence of about 20 consecutive heads. In almost any other context imaginable, 20 heads in a row would be incontrovertible proof that the coin is vitiated. But this coin is honest: the contrived setup made the amazing coincidence inevitable.
Even if the production of speculative epidemiology on an industrial scale were well-intentioned (and not, say, motivated by “publish or perish” anxieties or ideological animus), its most salient effects are reducing public trust in science and creating a vibrant quack market of “protection” against non-existent threats.