<p>I don't hear from "those on the inside" that the legacy preference is real at all. Quite the opposite. The legacy preference, if any, has to be distinguished from a developmental preference, which would benefit a small fraction of legacies and a few non-legacies. That exists, but is probably meaningless in terms of numbers.</p>
<p>I'll give you a few anecdotes from this year:</p>
<p>Stanford: East Coast boy, 2340 SATs, top 1% of public magnet school, good ECs with lots of leadership, one BA parent (very involved in alumni groups), one prof degree parent, >$900,000 cumulative giving</p>
<p>Harvard: girl, 2250 SATs, top 1% of famous suburban public high school, decent ECs, serious crew athlete, between the parents two BAs and two prof degrees, all summa, involved alumni, >$500,000 cumulative giving</p>
<p>Yale: boy, 2300 SATs, top-5 student at famous private school, decent ECs, very "pure" intellectual kid, one parent with BA and PhD, >$500,000 cumulative giving, kid accepted at Trinity Coll (Oxon.)</p>
<p>None of those kids was accepted. Non-legacy (and non-URM, not recruited athletes) classmates with equivalent (or lower) stats were accepted from each of the high schools involved. Tell me there's a meaningful legacy preference. (Of course, I could give you an equivalent list of legacy kids admitted, but the kids are really, really impressive -- as were the rejected kids.)</p>
<p>The larger point, though, is that you can take legacy admittees off the table as inconsistent with your top 1-2% idea. No legacy kids are getting in who are not in that group, unless they have something else really major going for them. </p>
<p>URMs don't account for enough of the student body at these schools to come close to 25% of a class, and many of them qualify in your top 1-2%, too. I'll give you another example from my son's school: URM, low SES, first generation high school graduate, 2200 SATs (non-native English speaker, 800 M), top 1% of class, top student of his ethnic group in the region, significant research ECs . . . and rejected at Harvard and Yale. I'm not going to say I think every URM kid admitted to those schools looked better in outline, but a bunch of them sure must have, and I don't think you can complain too much about them. And the larger point is that Harvard and Yale rejected this kid because they had 2,000 or so better applicants. If they took URMs or non-URMs with lower stats, it's because they had much better something else meaningful. This kid is your 1500/high GPA kid, and a URM to boot, and he was rejected (although I would guess he did have a 50% shot at admission). To me, that means that the people accepted instead of him are probably pretty darn impressive (the ones I know are).</p>
<p>Athletes: many of them would come within your top 40,000 group. The ones who don't amount to maybe 1-2 kids/sport/year, except for a few high profile/high body count sports like football and hockey. 100 to 120 kids per school per year, tops.</p>
<p>(And of course there's lots of opportunities to double- or triple-count among athletes, legacies, and URMs.)</p>
<p>So if almost everyone admitted comes from the top 2% pool, then why aren't the stats higher? Because the top 2% of the students did not get the top 2% SAT scores. That's all.</p>