Lastly, as far as large, nationally-recognized public schools go, they get a natural bump in admission stats because, unbeknownst to many who apply from out-of-state, most state schools are required to have a class composed of 75%-90% in-state residents (leading to much higher entrance requirements for 10%-25% of the class).
Um, shouldn't that lower
your stats? Because even though you have to be more selective with your out-of-state population, I would guess the in-state candidates you take have lower numbers than the out-of-state candidates you would replace them with if you operated like a private school without the requirement (if this weren't the case, then you wouldn't need the requirement). And ABA/US News look at all matriculants, not just out-of-staters.
Also, 75-90% seems high for top schools. I'm pretty sure at UVA it's 40%, and I'd imagine Michigan and the UCs are probably similar; I think Texas is a little higher.
ETA actual in-state numbers:
Boalt = 50% of acceptances are in-state to create a class that's majority residents (source
UVA = 40% of places are reserved for residents (source
Michigan = residency is a soft factor; incoming classes are 25% in-state (source
UCLA = for c/o 2010 it was 65% residents; no mention of a requirement (source
Texas = at most 35% can be non-residents (source
I should note that I'm not trying to prove a point here or anything, but I was just curious what these ratios looked like, and since I went to the trouble of looking them up, I thought others might be interested also. It's pretty off-topic, now that I look at it, although it might somewhat address JNSL's speculation about whether state schools have the maneuverability needed to really advance in the rankings. Anyway, sorry to derail, back to your regularly-scheduled Vandy thread...