NOTES
Whenever I hear correlations with implied causality I try to think of alternatives that fit the same correlations - sometimes the alternatives are implausible, but other times they are very plausible. The point is that the correlation doesn't distinguish between the two.