Newsflash: Correlation is Not a Cause!

July 13, 2011
159 Views

Just about every data scientist and statistician knows that correlation doesn’t necessarily confirm causation. However, popular business and social literature often confuse the two concepts. By understanding the maxim of “correlation is not a cause” more clearly, it’s possible to let loose creativity and imagination of the questioning mind.

Just about every data scientist and statistician knows that correlation doesn’t necessarily confirm causation. However, popular business and social literature often confuse the two concepts. By understanding the maxim of “correlation is not a cause” more clearly, it’s possible to let loose creativity and imagination of the questioning mind.

Humans like to think and speak in declarative terms. For a sampling, imagine statements bandied about by pundits such as; “global warming is caused by humans”, or “the Financial Crisis of 2008 was caused by greedy bankers” or “Republicans lost the 2008 election because they didn’t pay enough attention to immigration issues.”

Psychologist and author Sue Blackmore says the simple reminder that “correlation is not a cause” (CINAC), would improve just about everyone’s mental toolkit. Case in point, in an Edge.org article, she gives an example of people filling up a railway station as a scheduled train approaches.  She asks, “Did the people cause the train to arrive (A causes B)? Or did the train cause people to arrive (B causes A)?” The answer she says is they both depended on a railway timetable (C caused both A and B)!

Editor’s note: Paul Barsch is an employee of Teradata. Teradata is a sponsor of The Smart Data Collective.

In linear systems, cause and effect is much easier to pinpoint. However, the world around us is considered a complex system where there are often multiple variables pushing an outcome to occur. Nigel Goldenfeld, a professor of physics at University Illinois, sums it up best: “For every event that occurs, there are a multitude of possible causes, and the extent to which each contributes to the event is not clear. One might say there is a web of causation.”

And author Richard Bookstaber says that it’s a difficult search to pinpoint cause in complex systems especially because, “a change in one component can propagate through the system to lead to surprising and apparently disproportionate effect elsewhere, e.g. the famous “butterfly effect””.

The concept that in complex systems there is a “web of causation” may not sit right with some individuals, especially since newscasters, publishers, and even a fair portion of scientists prefer to insist on simple declarations of true and false.  However, the very nature of complex systems is that every object is in some way linked to another with either weak or strong ties and often connections are opaque and mysterious. So even a correlation of one, may not necessarily mean A causes B.

Freeing ourselves from the shackles of CINAC thinking means we have the possibility to let loose our imaginations says psychologist Sue Blackmore. Each definitive report should be greeted with skepticism. And when “A causes B” is held up as the answer, Blackmore cautions, then the critical mind automatically gets to work thinking; “Maybe instead, B actually causes A! And if not, what are the other opportunities?”

It’s human nature to try and explain the world around us. However, when it comes to complexity, we should lead discussions with a measure of humility to include questions and possibilities rather than declarations of certainty.

Questions:

  • How does our “aim to explain” end up stifling innovation and creativity?
  • The US Securities and Exchange Commission posited an explanation for the May 6, 2010 “Flash Crash”, but experts are not buying their simplistic explanation of a single “trigger event”. What are your thoughts?