Correlations Are Not Causation

In making plans for the authorization of charter schools in Washington state, we need to encourage policies founded upon solid research and discourage a common error in analyzing data.  Correlation does not prove causation. 

This reminds me of the first lesson my college statistics professor taught us.  This was back in the 60’s when polio was still a scourge and its cause only recently found.  He told the class that he had discovered the cause of polio.  He had graphed out the cycle of new incidents of polio and compared it to possible sources and found that the consumption of a soft drink called Bubble Up was the cause of polio.  He showed us his graphs and the rise in the incidents of polio matched the rise in consumption of Bubble Up exactly.  He pointed out there was a clear correlation between the two and concluded, therefore, that Bubble Up caused polio.

The absurdity of that conclusion would not be challenged by anyone in today’s world.  And that’s why it’s the perfect example to use to explain the difference between “correlation” and “causation”.  In today's educational reform efforts, too often correlation is presumed to be causation.