[danielf]

Further to Dr Wadd's comments: If two things are correlated, it does not mean they are causally related. As an example, there is a huge correlation between the number of churches and pubs in cities. That does not mean that churchgoers are heavy drinkers, but rather that you get more of both in bigger cities.

In the cannabis example it may well be that there are mediating (social) variables, such as getting both drugs from the same source (as the article acknowledges).

From a scientific point of view the article has only shown that the already established correlation between cannabis use and hard drug use does not appear to have a genetic basis. What the cause for the correlation is remains unclear.