Page 1 of 1

LSN Generalizability

Posted: Thu Dec 27, 2012 6:08 pm
by Teflon_Don
How representative is the LSN population to the actual law school applicant pool? For example, I see only 150-200 people listed for a school on LSN, when they admit 750-850 per year. And yet people still make huge statements based on those LSN numbers. Please explain

Re: LSN Generalizability

Posted: Thu Dec 27, 2012 6:25 pm
by onionz
Teflon_Don wrote:How representative is the LSN population to the actual law school applicant pool? For example, I see only 150-200 people listed for a school on LSN, when they admit 750-850 per year. And yet people still make huge statements based on those LSN numbers. Please explain
The bigger problem isn't the number of applicants, but the self selection of smarter kids using lsn generally.

That being said, the patterns for schools are really obvious after a few years, so that it becomes pretty reasonable to say what happens typically at each school. Those statements tend to be right.

Re: LSN Generalizability

Posted: Thu Dec 27, 2012 6:29 pm
by suralin
Regulus recently compiled some data and made a spreadsheet/tables showing LSN's representativeness to the applicant pool for specific top schools and in general--IIRC, pretty representative. I can't seem to find it right now though, so maybe try the search function or PM him.

Re: LSN Generalizability

Posted: Thu Dec 27, 2012 6:51 pm
by hume85
Teflon_Don wrote:How representative is the LSN population to the actual law school applicant pool? For example, I see only 150-200 people listed for a school on LSN, when they admit 750-850 per year. And yet people still make huge statements based on those LSN numbers. Please explain
A sample size of 150-200 is plenty big. There are other issues as others have/will mention.

Re: LSN Generalizability

Posted: Thu Dec 27, 2012 6:54 pm
by sinfiery
Suralin wrote:Regulus recently compiled some data and made a spreadsheet/tables showing LSN's representativeness to the applicant pool for specific top schools and in general--IIRC, pretty representative. I can't seem to find it right now though, so maybe try the search function or PM him.
https://docs.google.com/spreadsheet/pub ... SlE&gid=28

Re: LSN Generalizability

Posted: Thu Dec 27, 2012 7:17 pm
by jym_dawg
sinfiery wrote:
Suralin wrote:Regulus recently compiled some data and made a spreadsheet/tables showing LSN's representativeness to the applicant pool for specific top schools and in general--IIRC, pretty representative. I can't seem to find it right now though, so maybe try the search function or PM him.
https://docs.google.com/spreadsheet/pub ... SlE&gid=28
There's another spreadsheet from Regulus that shows the medians for a) current applicant pool and b) last year's applicant pool. Can't link to it now but you'll find it in any of the T14 threads. It shows that the typical LSN applicant is above each school's true median. The trend is more pronounced for some schools (e.g Duke's median for LSN applicants last year was 172 - 3 pts above the actual median) while others (UVa and Penn come to mind) are within 1 pt of the actual LSAT median. As others note, LSN is useful for observing trends and outcomes for people with similar numbers but generally LSN is not all that representative of the overall applicant pool. I imagine the difference from self-selection increases the further down the rankings you go.

Re: LSN Generalizability

Posted: Thu Dec 27, 2012 7:53 pm
by Teflon_Don
It seems very difficult to observe the other 25-40% of candidates that make up a school's admit pool, which is important for those who don't fall exactly at median