When we do SenseMaker™projects, one of the things that regularly crops up is the difference between correlations and causation. It's often the case that people read more into correlation numbers than is appropriate. And I always like having good examples for strong correlations that can be over-interpreted. Up until now, I've tended to fall back on divorce rate and national statistics.
But now, courtesy of the excellent Tim Harford (who's "Adapt" I'm enjoying as a well-written layman's guide to approaches to complex systems and environments), I've got a new one. I won't spoil the surprise - go read it here. The cheap and short version - size matters...and correlates.