that like saying you can increase tax revenue by cutting taxes
silly republican...
seriously though, i have played in almost a dozen live pokerrooms and i have played poker online at 15 sites 1000 hands or more. i am almost 100% sure the increased rake ISN'T compensated for by a better 'shark to fish' ratio.
honestly, a lot of losing players are aware of rake. a lot of winning players care a whole lot more about playing bad players than paying more in rake. also there are many factors other than rake to consider online (bonus, software, support, cashout times, interface isssues) and all those factors play a fairly big role in where the average player ends up playing, shark or not.
as a result there is a lot of noise that drowns out the 'high rake= massive fish' signal.
people talk about how absolute or gamesgrid were full of rocks and how pacific was full of fish. i honestly did not see a significant change in my winrate between the three sites (pre bonus and such...). weak tight players make mistakes just like loose passive players, they are just different mistakes. and there were tags at pacific and loose passives at gamesgrid and absolute too.
EDIT: anecdotal evidence: the average 2+2er is very aware of rake. the average 2+2er is FAAAR from being a shark. personally i hate the whole fish/shark paradigm dichotomy thing... but just to stay consistent with your terminology i would say at least a third(probably more) of 2+2ers are fish. furthermore, i would guess that at most 10-15% are 'sharks'