Following a statistic through several articles can lead straight back to another article, with no clear original study in sight. The wording sometimes changes along the way too.
Where do you stop chasing the source? Would you leave the number out if you could not verify what was actually measured?
.
