But this kind of cause and effect pattern actually does occurs. For example, there is a correlation between criminal activity and the mothers of the eventual criminal smoking during pregnancy. Now, it's possible that smoking during pregnancy somehow damages the brain of the child in a way that helps the child grow into a criminal. But it's also plausible that the kids turn into criminals for the same reasons that pushed their mothers to start smoking.
In general, alarms go off in my ahead when I read correlation as the justification for cause and effect. This is why I had a really hard time with social sciences. These sciences simply conduct experiments to find correlation between different variables, and simply assume that one variable causes the other.