Statistics – Destroyer of Superstitious Pretension

Statistics-education-research-day1 In Philip Ball’s Critical Mass: How One Thing Leads to Another, he articulates something rather profound: statistics destroys superstition. The idea, once expressed, is simple but does not stem its profundity. Incidents in small numbers sometimes become ‘miraculous’ only because they appear unique, within a context that fuels such thinking. Ball’s own example is Uri Geller: in the 1970’s, the self-proclaimed psychic stated he would stop the watches of several viewers. He, perhaps, twisted his face and furrowed his brow and all over America watches stopped. America, no doubt, turned into an exclamation mark of incredulity. What takes the incident out of the sphere of the miraculous, however, is the consideration of statistics: With so many millions of people watching, what was the likelihood of at least some people’s watches stopping anyway? What about all those watches that did not stop?

Our psychological make-up seeks a chain in disparate events. Our mind is a bridge-builder across chasms of unrelated incidents; a credulity stone-hopper, crouching at each juncture awaiting the next link in a chain of causality. To paraphrase David Hume, we tend to see armies in the clouds, faces in trees, ghosts in shadows, and god in pizza-slices.

Many incidents that people refer to as miraculous, supernatural, and so on, become trivial when placed within their proper context. Consider the implications of this: Nicholas Leblanc, a French chemist, committed suicide in 1806; Ludwig Boltzmann, the physicist who explained the ‘arrow of time’ and gave us the Boltzmann Constant, committed suicide in 1906; his successor, Paul Ehrenfest, also committed suicide, in 1933; the American chemist Wallace Hume Carothers, credited with inventing Nylon, killed himself in 1937. This seems to ‘imply’ a strong link between suicide and science. Of course, as Ball indicates himself, we must look at the contexts: We must ask what the suicide-rating of these different demographics was in general: of Americans, Europeans, males, and any other demographic.

Read more »