A colleague elsewhere has sent me some remarkable data, gleaned from ABA materials (for 2006-07), showing which schools are taking the largest number of transfers as a percentage of their first-year class. Let me preface this data by noting that 22.5% of the US News ranking of a school is based on the median LSAT and GPA of the 1L class; the smaller the class, of course, the easier it is to report better medians (one reason US News favors small schools, all things being equal, over large ones). (Note also that US News only counts the full-time day class, not part-time night students, though that's not at issue in the data below.) Of course, a school that reduces the size of its 1L class needs to make up the lost revenue, and one way to do that is by taking a large number of transfers. The transfers are "invisible" as far as 22.5% of the US News ranking is concerned, though their tuition dollars still pay the bills.
These are the ten schools which take the largest number of transfer students, relative to the size of their first-year class. With several of these schools, I strongly suspect the unusually large number of transfers is, indeed, explained by an attempt to "game" U.S. News. That being said, there can be other, more benign expectations for large numbers of transfers, such as bargain in-state tuitions. And even if these schools are trying to game US News, several of them are still wildly underranked (e.g., FSU, Rutgers-Camden, Illinois)!
1. Florida State University: 11 students transferred out, 59 transferred in, for a net gain of 48, or 24.62% of the 1L class.
2. Rutgers University, Camden: 4 students transferred out, 44 transferred in, for a net gain of 40, or 18.26% of the 1L class.
2. Washington University, St. Louis: 7 students transferred out, 51 students transferred in, for a net gain of 44 or 18.26% of the 1L class.
4. University of Illinois: 9 students transferred out, 37 students transferred in, for a net gain of 28 or 15.22% of the 1L class.
5. Georgetown University: 13 students transferred out, 100 students transferred in, for a net gain of 87 or 14.82% of the 1L class.
6. New York University: 0 students transferred out, 51 students transferred in, for a net gain of 51 or 11.49% of the 1L class.
7. Emory University: 6 students transferred out, 28 students transferred in, for a net gain of 22 or 10.63% of the 1L class.
8. University of California, Los Angeles: 3 students transferred out, 38 students transferred in, for a net gain of 35 students or 10.42% of the 1L class.
9. Northwestern University: 10 students transferred out, 34 students transferred in, for a net gain of 10.30% of the 1L class.
10. Columbia University: 1 student transferred out, 40 students transferred in, for a net gain of 39 students or 10.18% of the 1L class.
I must say these startling figures give me real pause about the reliability of this kind of measure of student body quality (quite apart from the limitations of LSAT as a measure).
For comparative purposes, consider that for other top schools, the more typical net percentage gain in transfers (relative to the size of the 1L class) is between 7-8% (Berkeley, Chicago, Michigan, Stanford, Virginia fall in this range), with Vanderbilt at the high end (9.47%) and Yale (4.71%), Duke (2.44%) and Texas (0.91%) at the low end, among others. Harvard had a net gain of 6.10%.
UPDATE: Sam Bagenstos (Wash U/St. Louis) writes:
I don't really have anything to say about the number of transfers we take - which has been roughly the same, as I understand it, since well before I got here - but there are a couple of points about transfers that may get lost in your discussion:
1. If, as you suggest (and I believe), the LSAT is a noisy measure of student quality and ability, it's entirely reasonable, wholly apart from US News considerations, for a school to defer a substantial number of admissions decisions until students can demonstrate their abilities during their first year at another school. I take it you agree that a school that followed that course wouldn't be "gaming" US News.
2. If a school were to try to "game" US News by taking large numbers of transfers who were less able, and less likely to succeed in law school, than the broad middle of its class, the effort would be inherently self-limiting, because US News counts placement as well (and, thinking more long-term, because students who can't get good jobs are unlikely to contribute to the school in the future). A smart school isn't going to take lots of transfers whose abilities can't enable them to get good jobs.
None of this is a reason to be unconcerned about the numbers you report, but it is a reason not to be too worried, I think.
There's something to both points, though I would add two comments. With respect to (1), this still leaves unexplained why some schools are taking much larger numbers of transfer students than others, who presumably have the same reasons to want to defer decisions on some students until they see evidence of success in law study. With respect to (2), job placement would be a meaningful check on reckless admission of transfers if the data schools reported to US News about placement was not essentially fictional; but the ways of gaming the placement data are legion and well-known. I have also heard anecdotal reports that some schools with no pre-screening for interviews are getting push-back from firms concerned about the large numbers of transfers these schools are taking.
I've opened comments now, if others want to add thoughts on this data. No anonymous postings will be approved; post only once, comments may take awhile to appear.
UPDATE: More thoughts on transfers from Bill Henderson (Indiana). I would urge everyone to read Professor Henderson's comments.
ANOTHER: More revealing data and analysis from Professor Henderson.