Rethinking cause and effect

The “cause vs. correlation” problem can be summarized like this: Just because A and B are related doesn’t mean that A caused B. We are often quick to assign meaning to any two pieces of data that seem to be connected.

A book called Spurious Correlations points out dozens of things that are correlated, but for which only a lunatic would believe there was a causal relationship. Among my favorites: It turns out that between 2000 and 2009 the divorce rate in Maine is perfectly correlated with the per capita consumption of margarine.

Before concluding that consuming margarine causes divorces in Maine, perhaps we can think to ourselves: Did A really cause B? Is there even a connection here?