Correlation, Causation, and Coincidence

I've been running around again. I'm back to CVille and to work, but before that I was out of town and away from my computer for almost a week, driving to New York, then flying to Georgia, with my "little" brother to get him set up for his new life in grad school. (I use the quotes because the 21-year-old is 6'2, more than a foot taller than me!) During my travel, I got a little reminder of one of my favorite logical fallacies:

Correlation does not necessarily imply causation.

Detecting baloney with the best of 'em... Carl Sagan.

This one comes in handy when reading that "studies show that more people with x do y." Such articles are usually written to give the impression that x causes y, when no such thing can be concluded from the work. This fallacy can also be used in an argument about anything, and it has of course been parodied in a few places.

So here is my example from my travels. On June 25th, I had two flights, one from Newark to Atlanta, and then from Atlanta to Valdosta. This was also the day on which Farrah Fawcett and Michael Jackson passed away. The news of each one broke while I was in-flight, so I learned of each when I turned my iPhone back on after landing. And I thought, well that was weird, two flights and two highly publicized celebrity deaths! Surely, strange but a coincidence. The next Sunday, I had the same flights back, in reverse. After the first flight, I learned that the news of Billy Mays's death had broken while I was in flight. At this point, it's safe to say that there was a correlation between my flights and celebrity deaths. However, is this a good reason for me to miss my next flight, as to prevent another one?

Well, no, I took the last flight (bumped to first class, mind you) without guilt, knowing that I would not be the cause of loss of life in that way. Just because two things occur simultaneously, without a plausible method for their connection, or other supporting evidence, there is no reason to think that one has caused the other, or even that they have a similar cause. (Do pirates expel less carbon dioxide? I didn't think so.) Well surely, this situation is an amazing coincidence? Okay, but for all of the possible coincidences that *could* occur to all people, all over the world, is it any surprise that a small fraction of them do happen? As a pattern-seeking species, we look for significance in random events, but sometimes we just have to accept that they are random. Think about that the next time you want to associate one seemingly random occurrence with another, or try to find meaning in life's coincidences.

For the record, the US soccer team did get beaten by Brazil right before take-off of our last flight. A small death for some? Only if looking too hard for patterns.

Cross-posted at One Astronomer's Noise