Quote from RealityBreak;978337:
Correlation doesn't need to imply causation for the causation to be there. There's a high correlation between getting your head ripped off and dying; I'm pretty sure there's some causation there too. It's pretty easy to correctly infer a cause from a correlation, even though statistics tells you you're not supposed to, because sometimes it's just blatantly obvious. (Gov. Health Insurance --> people live longer --> moar old people)
All causation demonstrates a relationship of correlation. Not all relationships of correlation demonstrate causation.
This is particularly the case when several correlated things are being caused by one or more unknown, unconsidered or incorrectly considered things. When one correlated thing is found to cause another, it is after it has been proven to do so, which makes that an isolated incident of cause and effect rather than a rule of thumb about correlation and causation, and words like majority, usually and most for that reason can only be taken with the same grain of salt as words such as some, rarely, and minority in a logical argument. For instance, your example (Gov. Health Insurance --> people live longer --> moar old people) shows a correlation of cause and effect because the causation has been
proven. Had you only demonstrated correlation and then proceeded to say that the correlation means causation ("people live longer --> Gov. Health Insurance --> moar old people" or "people live longer --> baby boom --> moar old people"), you'd be wrong.
That there would be a causation at all in your example owes itself more to the contextual relationship between the correlated things and not necessarily their innate correlation. In other words, whether or not the causation is there has nothing to do with the correlation. The correlation, on the other hand, has everything to do with the causation. Treating a logical fallacy as a rule of thumb, no matter how many times that logical fallacy coincidentally works out, will eventually lead to sloppy results. For instance, what would have happened if someone concluded that people get their heads ripped off when they die, and not the other way around?
In Europe half a millennium ago people didn't bathe regularly because they noticed a correlation between water and death. At best, establishing correlation works as a magnifying lens that can attempt to isolate the cause, but that's not a guarantee and your mileage will ultimately vary. You may even be focusing on an entirely wrong part of the puzzle, from which the cause cannot be readily discerned.