I read an interesting chapter in Ben Goldacre’s book Bad Science last night. He covered a question I’m sure many have answered before but in a more intuitive way. The question was: Why do we need science?

And the answer wasn’t because the world is complex that we need equations with eleven variables to describe the interactive population dynamics of lynxes and rabbits. Nope. His answer, and the right answer in my opinion, is that when humans look at the world we are prone to make two mistakes: 1) to see patterns where none exist and 2) to think that one things causes another just because one happened before another.

As an example of the first, many studies have been done where you give a person a random list of ones and zeros and many people will see a pattern. The example from the book as showing people a list of shots made by a pretend basketball player. The list was a randomly generated list of hits and misses yet many people see patterns in list–periods when they think the player is hot or not. Another example comes the Radiolab episode Stochasticity where the two hosts, Jad Abumrad and Robert Krulwich, were asked to draw up a random sequence of heads and tails. When you compare their sequence to an actual sequence of random coin tosses, the random sequence tends to have much longer strings of either all heads or all tails. Bottom line: we’re really bad at identifying randomness because our brains are always telling us there’s a pattern.

The second thing Goldacre talks about is something that is taken very seriously in epidemiology (and, I assume, in most other sciences as well). Most of the time, epidemiologists are only looking for associations between two things and only rarely do they ever have the gall to say something causes something else. On the news you hear, “X causes cancer,” but when you read the actual paper it says, “we’ve found an association between X and cancer.” Most, or many, or at least some, scientists recognize that it’s very difficult to say one thing causes another so we try and hold out until there is good evidence. When you start making too many connections between things that happen one after another you start getting into superstitions. Having played hockey my whole life, I know what that’s all about….