At some point you’ve likely heard the truism that “correlation does not imply causation”. It’s a phrase that reminds us that just because two data points line up, there isn’t necessarily a connection. Another phrase I’ve loved for many years is the latin phrase “post hoc ergo prompter hoc”, which literally means “after this, therefore because of this”, a logical fallacy that can easily be spotted in this example: “the rooster crows when the sun rises, therefore the rooster’s crowing causes the sun to rise”.
I’ve written a number of times about the biases and mental tics that lead us to bad conclusions. Here is a post that references Nassim Nicholas Taleb’s “Black Swan” concept, and here is one of many posts on cognitive biases.
Sometimes the best way to really learn an idea is by laughing at the ridiculous, so it was with great joy that I ran across Tyler Vigen’s excellent website called, appropriately enough, Spurious Correlations. Tyler created an engine where clearly unrelated data sets that mimic one another can spit out a graph that makes you wonder if just maybe there might be some connection. Here are a couple great examples from his website….
We live in an age of opinions masquerading as facts. Making a critical business decision based upon a spurious correlation (that is likely not as ridiculous as the above examples and thus difficult to spot) can be suicide. Speaking of suicide….
Are you questioning correlations that you accept as implicit? What practices do you use to get away from the fire-hose of emails you react to and think about what correlations in your industry are not spurious but are actually driven by causation?
The need for crisp, logical, inquisitive thinking has never been greater.