A recent article in Natural History magazine does not exactly advise eating dirt, but rather examines what the author calls “the hygiene hypothesis,” that is, the hypothesis that the increasing prevalence of asthma and other autoimmune diseases is the result of excessive cleanliness. Along the way, the author, Druin Burch, illuminates in some detail just how good science works.
The gist of the hygiene hypothesis is that the more exposure people get to the germs and parasites with which they coevolved, the less likely they are to develop autoimmune diseases like hayfever. The prevalence of allergic and autoimmune diseases has increased in the last 50 years or more, and, moreover, such diseases are more common among “the richer citizens of the world, [for whom] exposure to many common microorganisms is being delayed.” Indeed, children who go to nursery school or grow up on farms are less likely to contract allergies or autoimmune diseases. It is possible that, paradoxically, an immune system that does not have enough to do looks for something to attack; sometimes that something is an allergen, and sometimes is the host’s own body.
Burch notes, however, that correlation does not prove causation, though he does not express it exactly that way. He argues,
There have certainly been lots of articles on the subject, including lots of peer-reviewed studies that seek to look at original data. I hesitate to term them all “science,” since many are descriptive: they design and carry out no experiment and attempt to falsify no hypothesis. They fall into the sort of technically proficient studies of natural history that most of us think of as science but which are really just the groundwork for it. They develop theories about the way the world works; science only happens when those theories get properly tested.
And later on,
The most important thing to note, though, is that what started off inconclusive has stayed that way.The association between infections and allergies was first suggested as long as half a century ago–although originally the idea was that infections caused such diseases, not prevented them. Allergies and autoimmune diseases have risen and parasites have dropped away at the same time as a host of other things have changed–democracy has spread, satellites have been launched, mobile phones have been developed, and a raft of modern politicians elected. If you wanted, you could show an association between any of those aspects of contemporary life and the increase in rates of allergies and autoimmune disorders.
Associations are easy to spot, even plausible ones–but proving causation is hard. Might early exposure to bugs and parasites be entirely unrelated to allergies and autoimmune diseases? Quite possibly. It’s been suggested that the spread of detergents in the world (which lines up chronologically with every other modern phenomenon) disrupts the mucus lining the inside of our intestines, with some diseases being caused as a result.
And finally he confesses that he “had the strong impression that research into the hygiene hypothesis had shown that dirt was good for us,” but found that he was “mistaken”; the evidence is “clearly weak.” Possibly, he had that impression because “[t]he hygiene hypothesis is the sort of story that’s attractive to mass media”; I certainly had that impression for that reason (though the hypothesis seems somewhat stronger to me than to Burch).
I cannot entirely agree that descriptive studies that formulate no hypotheses are not science; if that were so, then taxonomy might not be considered a science. But it is important that 50 years of careful epidemiological work is not conclusive. Of more concern to me here, however, is the fact that the very scientists who work on the hygiene hypothesis remain unconvinced. Specifically, they do not formulate a hypothesis and then break their necks to support it. Rather, they look at the evidence as dispassionately as anyone can, and they stand ready to discard the hypothesis if it is falsified.
Such behavior is in stark contrast to any of several ideologies I could name.