Law School Discussion

"Splitter" questions

Re: "Splitter" questions
« Reply #10 on: October 08, 2005, 07:58:03 AM »
It seems to me that splitters wont have any necessary disadvantage over anyone else with one score over the median and one score under.  Yes schools now dont have any incentive to pull up their 75 score so dont need to inflate it by accepting those 3.2/176ers but still will need people above the medians. For a school like Georgetown with an LSAT median of 169, there are a lot fewer 170+ people than there are people above their GPA median (not sure what it is maybe 3.6 or 3.7). Those 170+ folks who also have GPAs that are above most T14 medians will more than likely be accepted at one of HYS and if not definitely CCN. Therefore Georgetown still has a big incentive to accept every 170+ person they can find as they need to keep that LSAT median up. The incentive to accept the person with the 176 versus the 170 is now gone though.

Re: "Splitter" questions
« Reply #11 on: October 08, 2005, 12:57:43 PM »
Another strange possible scenario...Schools will be throwing scholarships at applicants that fall within or slightly above their medians, as these people will maintain/improve that median. Theres no longer a reason to give a 3.9, 176 applicant a big scholarship. Let him go to Harvard, theyll say. There are a limited number of students out there that maintain/improve the 7-14 median, and an applicant with a 3.9, 176 isnt one of them.

Re: "Splitter" questions
« Reply #12 on: October 08, 2005, 03:25:10 PM »
The incentive to accept the person with the 176 versus the 170 is now gone though.

The incentive for places like GULC, NW etc. is to maintain their current median - i.e. admit every 3.5-3.8, 167-171 they can find. I think youll find more even score distributions across the board. Why accept a 176 that doesnt maintain your numbers when you can accept a 170 that does? Its not so much that the advantage of a 176 has been lost as it is that its become a liability to some degree. In the top 6, its still going to serve you well. Outside that, though, schools have no reason to have high 75th's anymore. The 75th might as well be the same as the median - it doesnt matter. On the 25th end (equally as meaningless), though, it frees them up to take more otherwise ineligible candidates with some redeeming qualities. As long as a certain percentage of a schools admits are within range of their target median, they can do whatever the hell the want with the rest of them. I think a lot of people on the high end are going to be sacrificed in the name of diversity. We'll have to wait until next year, but I'd bet youll see schools patting themselves on the back for their increased minority enrollment while their 75th's drop and their medians stay the same.

I just dont agree that people on the high end are going to be sacrificed in the name of diversity at least as it relates T14 schools. If you do a search of people who applied to schools in the 7-14 range on lawschoolnumbers over the past two years most of the people who applied with both numbers around the median were already being accepted under the 25/75 system. I think the people who will lose out the most under this new median system are those who are not URMs or have other "diversity" factors and are under both medians, for instance between the 25% and the median. I think these people are really going to suffer. How many people with 169s or 170s and 3.6-3.7 gpas were not already being accepted to most 7-14 in addition to one or more T6 over the past several years. I would guess not very many.


Re: "Splitter" questions
« Reply #13 on: October 08, 2005, 03:41:55 PM »
His entire argument is pretty convoluted.  And his math is wrong.


  • ****
  • 1053
  • Hey psycho stalker spy!
    • View Profile
Re: "Splitter" questions
« Reply #14 on: October 08, 2005, 04:01:22 PM »
I'm an English major, so my only contribution to this discussion is, what does this 25-75 v. median change do to my chances? I have about a 3.45 and probably a low 170 score. Are my chances worsened or bettered?


Re: "Splitter" questions
« Reply #15 on: October 08, 2005, 04:02:54 PM »
Very curious whether a mega-split of something like 2.3, low to mid 170s would be affected by this.... or if I'd just be looking at tier 4s anyway.


  • ****
  • 316
  • "No way, man." Phrase: to have disbelief.
    • View Profile
Re: "Splitter" questions
« Reply #16 on: October 08, 2005, 04:59:21 PM »
Honestly, I think it will help you.  Your 2.2 will not bring down their gpa, while your lsat will be above *most* medians.  High Lsats are harder to find than high gpas, and with the new system they can very easily hide low numbers.

Re: "Splitter" questions
« Reply #17 on: October 09, 2005, 02:41:01 PM »
Heres a mathmatical demonstration.

170, 166, 167, 167, 168, 173, 166, 169, 170, 168, 171 = Median of 168

Throw in some new numbers and see what happens
180 = 168.5
169 = 168.5

Using the mean instead, the mean is a 168.6.
180 = 169.58 (+ 1)
169 = 168.66 (+ .1)

Using the median, the 180 is no better than the 169. If the 169 applicant is more interesting, theyll probably take him. Essentially, this means that in the top 14, a 180 isnt going to give you an advantage over the average admit. Also, there is no reason for them to give the 180 a big scholarship when they can give half as much to a 169 (who will probably be greatful for anything) and still retain their median. The 180 still looks impressive, but schools wont feel compelled to take/persuade you like they would have under the mean based rankings.

On the bottom end, you can literally put in a 120, and the median is still a 168. In fact, you can add five 120's before you see any effect on the median at all - even then, its just a 1 point decrease. Nearly a third of your scores can be 120's, and you can still have a 168 median. On the bottom end, this makes any given individual with a 120 no different than one with 168. Obviously, they cant do this in every case. But if youre a URM with a 120, theyd be better off taking you over some kid with a 168. If the top schools want to horde all the URM's in the country, they can do so without affecting their numbers. Likewise, if you fall short on one median but bring the other one up, youre better in a heads up comparison than a candidate who hits the median on both. With means this isnt necesarily true.

The schools will have to accept a certain amount of students around their target median. After that, theyre going to have a lot of spots left to take whoever they want to, regardless of stats. Average applicants will be the most sought after (the top schools first priority is in maintaining their median, and since theyve all got similar medians, theyre all fighting for the same applicants. The battle will be fought with scholarships.), top applicants will lose some ground (theyll still get into HYSCCN, but the other top schools have no reason to give them lots of money. Theyre no different than barely above average applicants in the t14), and splitters have a much, much better chance than they did when rankings were done with means.

Again, this all assumes that USNWR is really running the show. If adcoms really believe in the admissions process, theyre still going to throw lots of money at the 4.0, 180 kids. Every school wants to have fantastic alumni. All Im saying is that with regards to the rankings, the schools have A LOT more freedom in selecting their students, and I think their repsonse to using medians is predictable.


Re: "Splitter" questions
« Reply #18 on: October 09, 2005, 03:51:43 PM »
Also, as long as were all assuming that USNWR runs this whole thing, theyre switching to medians rather than means. This will things very interesting for splitters.

All right, I'm gonna take issue with this.  For the 2006 rankings (influenced by the 2003-2004 cycle) U.S. News changed their methodology.  They began calculating their own midpoint, versus using a school's self-reported median.  They calculated this midpoint by simply averaging a schools 25th/75th percentiles.  U.S. News did this in response to what they were believed were questionable self reported medians (50%). The ABA releases actual 25% and 75% of it's member schools.  So, in fact last year was the first time a mean entered into the USN&WR rankings.   

U.S. News, operating under the assumption that law schools will not fudge their ABA data and risk losing accreditation, simply took the ABA numbers.  Last year was the first time they used this method. Leiter has pointed out several flaws with this method (as he does with every U.S. N&WR method).  However, I haven't found any articles stating that the ABA will release member school medians to U.S.N&WR.  If they are out there, could someone please post a link.  I also have been unable to find any information stating what the U.S.N&W.R. methodology will be for the 2007 rankings.  I am aware that the fact that I haven't found the information does not preclude it's existence.  In fact, I would appreciate help locating it. 

Just remember, the rankings are meaningless according to the ABA. And they are, at least until law schools begin paying attention to them to the degree that they change their operations.  I'm actually going to start a new topic about this.  I want to hear hard facts regarding an announced change (not XOXO or LSD gossip).  I'm taking this seriously, because it has a huge effect upon my application strategy.

Please Comment.

Re: "Splitter" questions
« Reply #19 on: October 09, 2005, 05:04:58 PM »
From Leiter:

U.S. News has now announced that it will go back to using the medians, but the ABA will also begin collecting and publishing the median data as well.  So, on the theory that schools that will lie to U.S. News won't lie to the ABA, the median data that goes into the U.S. News ranking is likely to be reliable.

As for 2007 policies, who knows. They made the change for this year on fairly short notice.

What are your stats and whats your strategy? The only real strategy change I would suggest based on this whole thing is that if youre a low GPA/high LSAT splitter, I'd apply to the schools you thought you were excluded from. According to this, even a place like Harvard could afford to take a 3.3, 178. Whether they actually will, who knows. I think theres enough confidence in the LSAT that they wouldnt do the same for someone with a 4.0 and 163. To the extent that this is true, it will be the high LSAT splitters that benefit the most.