Well I said I didn't read most of the thread. xP
Though I still think that "can cause" is misleading. There is such a thing as too large of a sample size, and with a sample that large you could find a correlation between pretty much anything. There are way too many variables in that situation.
