Correlation is not causation – but it’s not not-causation

“How today’s dope apologists are just like Big Tobacco apologists 60 years ago.”


From Peter Hitchens of the Daily Mail

You can’t blame the Cannabis Comment Warriors for ceaselessly pointing out that ‘correlation isn’t causation’ . They are desperate.

For they are in much the same position that Big Tobacco was in when it first became clear that cigarette smokers were becoming ill in large numbers. The correlation was there. The cause couldn’t be proved. So, rather than admitting the risk, they reached for the most complacent possible answer, so delaying the general recognition of the truth for many years.

The modern campaigners for Big Dope – tantalisingly close as they are to legalization – fear this topic greatly. They know very well a) that many marijuana users become mentally ill , b) that mental illness itself is extremely hard to categorize objectively and c) that the mechanism by which they become ill is currently impossible to discover, and may remain so for many years. The same was true, for a long time, of the undoubted correlation between cigarette smoking and lung cancer, and the same claim was emplyed by Big Tobacco as is now being used by the people who hope to be Big Dope.

They are quite right to say that correlation is not necessarily evidence of causation. But they are less keen to admit the corollary.

Correlation is not necessarily *not* evidence of causation, and in fact is one of the foundatons of epidemiology.