A
problem encountered in
statistical analysis when your two
groups (or however many you have) both have such high
scores on whatever
measurement you use that you couldn't possibly tell the
difference between them even if there was one. This problem occurs because statistical
tests compare two
populations by comparing their
means on whatever measurement you choose, and seeing if the
variances of the
samples are large enough that they might both be from the same
population. So, if all of your scores are really high, then your means will be very similar and so you might wash out any effect that there might be. The solution is generally to use a more difficult test.
This problem has a potentially interesting analogy the E2: if a certain type of node tends to get upvoted a lot, then you might be unable to tell which of those of nodes are the really good nodes, because all of them will have high scores.
And I admit that I added the previous paragraph in part because of the noted effect of nodes about noding getting more upvotes.