It seems like it could be somewhere in between. If Chicago really is interviewing 50% of applications, the admit after interview rate is probably not as high as 50% but it's also probably not 25%. That sounds like a lowball figure except if Chicago is hardcore utilizing the waitlist to manage yield or had an application spike even higher than its undergraduate program did over the past two years.
Using the BW figures published last year, which I think are off class of 2011 profile stats (? If they're off 2010 instead it wouldn't alter the numbers too too much) and the Booth website (Rhyme, shoot me down if I'm making stuff up), I don't think the admit after interview was as high as 50% last year.
592 students in Class of 2011 (Booth website) / 3843 applications (BW) ~ 15% enrollment rate
BW is claiming 22% selectivity from their chart, which would imply ~ 845 admits.
592 students / ~845 admitted students ~ 70% yield, which seems high but I guess that was the thing last year? Unexpectedly higher enrollment?
If Booth interviews half of applications, that's ~ 1922 interviews.
~ 845 admitted students / ~ 1922 interviews = ~ 44 % acceptance rate after interview -- including waitlisted students that were later admitted. So true admits would be somewhat lower and if you buy that Round 1 is maybe easier than Round 2, perhaps a couple of degrees lower still.
So now I'm thinking, what if you assume that Chicago experiences as crazy an application spike as their undergrad program, still interviews half of applicants (applicant pool would have to not be watered down at all, despite the extra apps) and has no plans to raise its enrollment for the Class of 2012. That's pretty much the only way you get close to a 1 in 4 figure.
40% spike in applications ~ 5380 apps
~ 575 target enrollment / ~ 5380 apps
~ 5380 apps * 50% IV rate ~ 2690 IVs
assuming same 70% yield off ~ 575 target enrollment ~821 admits.
~ 821 admits / ~ 2690 IVs ~ 30% acceptance rate after the interview -- which would again include waitlisted applicants that are later admitted and any adjustment you want to make R1 v R2.
So assuming all those crazy assumptions about volume, quality and adcom decision-making, its possible she's not lying but it seems less likely than a slightly higher number.
It's also possible I made all this up so you should ignore it.