Joe's point, i think, is you can't use the totality of the data. you need some sort of defining criteria to separate and analyze. i don't care what it is, faves of 101 to 120. faves of 121 to 140....