The Sun today published an article about the connection between smoking parents and delinquent children. A survey asked 830 teenagers about their parents’ smoking habits and their own criminal activity. Those who reported that their parents smoked were statistically more likely to admit to going off the rails.

Professor Colin Pritchard, of Bournemouth University, offered his interpretation of the data. The conclusion that he drew is that there is a causal connection, that parents smoking causes their children to misbehave. He therefore called for new warnings to be displayed on cigarette packets, detailing the social and psychological effects on one’s offspring that smoking can have.

Pritchard’s conclusion is over-drawn. The mere fact that there is a correlation between parents smoking and children misbehaving does not imply that one causes the other. To suggest that it does is to commit the correlation not causation fallacy.

Besides, if we must conclude that there is a causal connection, why not infer that it is in the opposite direction, that having delinquent children causes parents to smoke? The correlation uncovered by the survey is equally good evidence for this conclusion as it is for that drawn by Pritchard.

In fact, though, it seems more likely that there is no direct causal connection between smoking parents and delinquency; it may well be that the two are independent but have a common cause, such as class.

Working class parents are more likely to smoke. Working class children have higher levels of delinquency. We would therefore expect to find these two things together even if there is no direct causal link between them.

A nice illustration of a fallacy from The Simpsons:

After a single bear wandering into town has drawn an over-reaction from the residents of Springfield, Homer stands outside his house and muses, “Not a bear in sight. The Bear Patrol is working like a charm!”

Lisa sees through his reasoning: “That’s specious reasoning, dad.” Homer, misunderstanding the word “specious”, thanks her for the compliment.

Optimistically, she tries to explain the error in his argument: “By your logic, I could claim that this rock keeps tigers away.” Homer is confused: “Hmm; how does it work?” Lisa: “It doesn’t work; it’s just a stupid rock!” Homer: “Uh-huh.” Lisa: “… but I don’t see any tigers around, do you?”

Homer, after a moment’s thought: “Lisa, I want to buy your rock…”

Correlation does not imply causation. Just because two things occur together, does not mean that one caused the other. Homer argues that as the Bear Patrol vans are correlated with an absence of bears, the former must have caused the latter. Lisa, tongue in cheek, argues that as the presence of her rock is correlated with an absence of tigers, the former must have caused the latter.

At least Homer recognises that the two arguments are on a par, even if he fails to recognise that both are examples of the correlation not causation fallacy.