And besides, it doesn't always correlate
If
there ever was one bumper sticker argument that I would like to see
eradicated completely, it would certainly be "Correlation is not
causality." This piece of pseudo-wisdom pops up constantly whenever
somebody points out a strong correlation that supports his argument,
and for every incident where this is a valid counter, there are
probably ten incidents where it is just stupid. However, since only a
few chosen people get to directly observe causality in action, while
the rest of us have to settle for merely observing correlations, the
argument by observed correlation will necessarily remain with us for a
long time. And each time the response will be "Correlation is not
causality, correlation is not causality! Bwaak! Correlation is not
causality! And besides, we must always remember the exceptions such as
[insert some emotionally appealing counterexample]" (thanks to Markku for summing up the common argument in the last sentence in such a concise form)
If it is impossible to eradicate this argument, at least it should be possible to counter it with some similarly effective bumper sticker. For this purpose, I propose "Correlation demands an explanation." It would work on so many levels at once.
If it is impossible to eradicate this argument, at least it should be possible to counter it with some similarly effective bumper sticker. For this purpose, I propose "Correlation demands an explanation." It would work on so many levels at once.
Comments