Causation, Correlation, and Ice Cream Sundaes

My sense of cause and effect in day-to-day situations is better now than when I was a child, but it's still pretty sketchy. Hey Dora, if you put a greasy potato pancake in the toaster oven naked it will drip oil on the element and cause a fire (oooooo is that why the house is all smokey?) --yup that happened a few weeks ago. And yup, that really is better than when I was a child and thought that whether I'd have a good day or not was dependent on what color "fireworks" I saw when I pressed on my eyes in the morning, and everything else "just happened," like weather.

Of course there is a causal relationship between flaming oil and smoke (i.e. flaming oil causes smoke to occur), even if I kind of missed it at the time. And of course there is no relationship of any sort between the "fireworks" and having a good day. But rarely in the world are the relationships so clear. This is because the world is fantastically complex, full of hidden variables, and much of what happens is highly correlated but not directly causal.

Fact: Ice cream sales are highest on certain days of the year.
Fact: Murder rates are highest on certain days of the year.
Fact: Ice cream sales and murder rates are often highest on the same days of the year!
Conclusion: Eating ice cream causes murder!

Analysis: Well, no, not quite. High temperature is really a cause of both the rise in ice cream sales and the rise in murder rates. There is definitely a relationship between the ice cream sales and murder rates though, it's just not directly causal; rising ice cream sales and rising murder rates are highly correlated through their common (but hidden) causal factor temperature. While correlation does imply that two variables are related (and that may even be, though not in this case, ultimately causal) it does not automatically mean that one variable must be directly determining the behavior of the other variable.

Further, just because temperature is "a" cause for ice cream sales and murder rates does not mean temperature is the only cause. Other factors such as availability of ice cream, availability of leisure time, and cultural norms regarding the eating of ice cream are just a few other factors besides temperatures that may "cause" trends in ice cream consumption.

When considering any causal type claims related to autism, be they something like watching TV causes autism, or something like a therapy causes an improvement in a child, it is crucial to question whether the relationship is in fact causal, whether it is merely correlated (potentially causal, but not necessarily--there may be other reasons for the apparent relationship!), or whether it is purely coincidental (not really related at all). It is also crucial to consider the possibility of more than one cause for any given phenomena. The media loves to make strong statements about X causes Y; however, science and nature are rarely able to provide that same degree of certainty.