Monday, July 03, 2006

Florida Explains the Data it Reported for the USN&WR Rankings

The most recent U.S. News & World Report law school rankings apparently err in favor of the University of Florida (Levin) College of Law. As explained in an earlier post, USN&WR evidently ranked Florida based on the median LSAT and GPA of only that school's fall 2005 full-time, first-year class. But USN&WR almost certainly intended to rank the school based on the (somewhat lower) median LSAT and GPA of its entire 2005 full-time, first-year class. It seems, in other words, that Florida benefited from the same sort of error that benefited Baylor University School of Law.

Did USN&WR in fact rank Florida based on the wrong median LSAT and GPA? And, if so, why? As I did in Baylor's case, I sought explanations from both USN&WR and Florida. I here offer their responses.

Florida's Associate Dean, J. Patrick Shannon, responded to my inquiry by replying, "Our USN&WR data is based on Fall admits only, per an agreement reached several years ago. The ABA [data] is based on both Fall and Spring entrance."

USN&WR's Director of Data Research, Robert J. Morse, countered with a different take: "Any arrangement or understanding that may have been made in the past is now null and void. The LSAT and GPA should be full-year class data, not fall only."

Florida's Dean, Robert H. Jerry, II, followed up with a letter more fully explaining his school's practices. I here offer a fair chunk of it, and will link to the full text [update: here it is] if Florida posts it online:
For virtually all of our law school’s modern history, we have admitted two classes annually from two separate application pools. This practice began in the mid-1960s, long before the USN&WR rankings appeared on the scene. Each class had approximately 200 students, with one matriculating in January and the other in August. Applicants admitted to one entering class were not moved to the other in the absence of the admitted student petitioning for deferral of the student’s start date. . . . There will be no spring 2007 entering class, and in future years we will have only one fall entering class per year.

In the late 1990s, the then-dean of the college sought guidance from USN&WR as to how Florida should report its entering class data given the fact of two entering classes per year. The college was advised by USN&WR to report data for the fall class only. Only a couple of administrators from this era are still on the Florida administrative staff, but their recollection, which is confirmed in our files, is that the then-dean had reached an agreement with USN&WR that fall-only data should be used in responding to the USN&WR request for GPA and LSAT information. Since that time, the college has consistently reported data for the fall entering class to USN&WR. . . .

At no time prior to June 26, 2006 did USN&WR inform us that the earlier instruction received by the then-dean had become invalid or been superseded by revised instructions. It is important to note that Florida has consistently reported merged data to the ABA since that data was first requested, and that information has always been a matter of public record in each year’s ABA-LSAC Official Guide to ABA-Approved Law Schools. At no time did USN&WR bring to our attention a difference between the data in the Official Guide and that published in U.S. News. This difference has always been obvious, and it would have been fair for Florida’s staff to assume, if they were to have given it this much thought, that USN&WR would notify Florida if there were a problem with Florida’s submission. In this regard, I must stress that the Florida staff who submitted data to USN&WR through the years did so in good faith and without any intent to supply inaccurate information to anyone.

In a June 26, 2006 e-mail from Robert Morse at USN&WR, we were informed that US News wants us to report merged fall-spring GPA and LSAT data. We surmise from this instruction that USN&WR wanted us to report the data in this manner earlier. Had we received this clarification earlier, we would have complied with this instruction earlier; if it were our plan to continue to admit two classes annually, we would comply with this instruction immediately. For future submissions to USN&WR, the revised guidance is moot, because, as explained above, beginning in the fall 2006 semester we will matriculate one class of 400 students in the fall of each year.

The claim in that last line left me wondering. Assuming, as seems likely, that USN&WR's fall 2006 questionnaire repeats the language used in its fall 2005 questionnaire, Florida will have to report the combined median LSATs and GPAs of its spring 2006 and fall 2006 full-time, first-year matriculants. The questionnaire looks backwards, grouping all of the prior year's matriculants for reporting purposes. On that analysis, contrary to Dean Jerry's understanding, the issue won't be moot for Florida until the fall of 2007. I emailed Dean Jerry to that effect. He replied, cc'ing Robert Morse, saying he would seek clarification from the latter. Morse replied, making it absolutely clear (to my mind, at any rate) that Florida would have to include the LSATs and GPAs of its spring 2006 full-time matriculants with those of its fall 2006 full-time matriculants. (Long-time readers may recall that the body of and comments to an earlier post discussed this same issue, albeit less conclusively.)

Dean Jerry and I do agree, however, in assuming that USN&WR intended to rank Florida based on the median LSAT and GPA of its entire 2005 full-time, first-year class. Why need we assume as much? Because, strictly speaking, Morse's terse denial, quoted above, used only the present tense. It did not expressly claim that any agreement between USN&WR and Florida was "null and void" when the school reported its numbers last fall, nor that the LSAT and GPA should have been "full-year class data." All the available evidence confirms that Morse did mean to complain that Florida had reported the wrong numbers, however. By way of confirmation, I explained my assumption to Morse in a subsequent email and invited him to correct me if I'd erred. He has not done so.

Dean Jerry's letter observes that USN&WR might have compared the data it got from Florida with the data, reported in the ABA-LSAC Official Guide to ABA-Approved Law Schools, that the ABA got from the school. Some weeks ago, when I was first looking into discrepancies between USN&WR's data and the ABA's data, I asked Morse about that option. He replied that the ABA releases its data too late for USN&WR to use in double-checking its data prior to the publication of its rankings. That is not to say that USN&WR couldn't use the ABA data to correct the published rankings, of course. USN&WR has shown no inclination to correct or annotate its extant rankings, however.

I close with one other bit of salient data from Dean Jerry's letter: He noted that one of my earlier posts inaccurately stated that Florida's fall 2005 full-time first-year class had a median LSAT of 156. The correct number is 159. I've now corrected that typo. (I need not correct any related results from my model of the rankings, as it did not suffer a similar error.)

Earlier posts about the 2007 USN&WR law school rankings:


duke.glosgow said...

Nice Blog Tom.This years rankings seems to see a notable change and it would depend upon the employment of the Scholar graduate. I would also suggest that the students and professors votes must also play a role in the final rankings.

Duke Glasgow
Florida Drug Rehab
Florida Drug Rehab

Yoga said...

I don't understand those rankings