If it's worth saying, but not worth its own post (even in Discussion), then it goes here.
Read the comments and weep-- it's almost as though there are a lot of people who resent specificity.
I wouldn't have been surprised if there were people who said "but randomizing means that half the people aren't getting the obviously valuable help!", but nobody said that.
A person A uses "common sense" to suggest a solution X, which seems obvious to them.
Research shows that in reality, Y is much better than X, and sometimes X is actually not even helpful.
A person B says: See? This is why you should always collect and evaluate data!
A person C says: No, the real lesson is that we should use common sense which obviously recommends Y! Why didn't anybody use their common sense?