How Low Will They Go? Forum

[nonexpostfacto]

I was searching through Law School Numbers today, and I was trying to guess GPA floors of the T14 for high LSATs. Even briefly looking at the numbers, it was easy to see that they've started to dip lower for high LSATs (Spliters Rejoice).

Will this trend hold in the next few years? How low will GPAs go in the future for a 175-180 score? My (really rough) estimations are below, given LSN. Is it fairly accurate?

[This is just a post out of curiosity. I haven't taken the LSAT yet, but have already started studying, probably for December.]

GPA Floor for a 175-180 Score
Yale: Doesn't Really Apply... But Around 3.8
Stanford: Doesn't Really Apply...
Harvard: 3.65
Chicago: 3.6
Columbia: 3.3 - 3.4
NYU: 3.2 - 3.3
Penn: 3.2 (Weird with Yield Protect/Around 3.0 with ED)
UVA: 3.2 (Weird with Yield Protect/Around 3.0 with ED)
Berkeley: 3.65 - 3.7
Michigan: 3.2
Duke: 3.3 - 3.4
Northwestern*: <3.0
Cornell: 3.1
Georgetown: 2.8

*ED at Northwestern doesn't seem to have a huge impact on splitter admission, and the full ride scholarship almost never goes to a splitter. Applicants with a high GPA [3.8] are much more likely to receive the full ride ED, with a 170-174. All of this is just an observation from LSN, though.

[RodneyRuxin]

I like your title and it seems like you did your research. I haven't looked into this, but with the frequent "retake and ED UVA" advice on TLS I'm guessing that is a factor you might want to consider.

[nonexpostfacto]

I like your title and it seems like you did your research. I haven't looked into this, but with the frequent "retake and ED UVA" advice on TLS I'm guessing that is a factor you might want to consider.

Good idea. I didn't check to see how many of the admits were applying ED. I'll take a look right now and see if I should add a bit more info.

[ralph]

For Yale:
https://officialguide.lsac.org/Release/ ... px?sid=177

The floor is pretty much >= 3.75 for the class of 2015(?).

[nonexpostfacto]

For Yale:
https://officialguide.lsac.org/Release/ ... px?sid=177

The floor is pretty much >= 3.75 for the class of 2015(?).

Yeah, looking at LSN again, I would say that a 3.8 might be a good guess for Yale... Even though there are still plenty of denials above that.

And Stanford is probably a 3.9, even though there are some admits with much lower GPAs (non-URM, too), so I don't think that there is a good floor to put for Stanford.

[Ti Malice]

*ED at Northwestern doesn't seem to have a huge impact on splitter admission, and the full ride scholarship almost never goes to a splitter. Applicants with a high GPA [3.8] are much more likely to receive the full ride ED, with a 170-174.

Yeah, NU is not targeting splitters at all with its ED option. NU is mostly going after CCN admits.

Yeah, looking at LSN again, I would say that a 3.8 might be a good guess for Yale... Even though there are still plenty of denials above that.

And Stanford is probably a 3.9, even though there are some admits with much lower GPAs (non-URM, too), so I don't think that there is a good floor to put for Stanford.

YLS and SLS both have pretty hard 3.8 GPA floors for non-URMs. Here are LSN outcomes at both for applicants with a 3.8-3.85 GPA and a 175-180 LSAT over the last several cycles.



And here are outcomes for applicants with the same LSAT range and GPAs from 3.7-3.79:

[Monochromatic Oeuvre]

Attempting to establish a numerical value on "GPA floor" requires a more precise definition of the term. Rarely do we use the phrase to literally mean "the lowest GPA at which a school will accept an applicant." For example, a non-URM with a 3.1 was accepted to Yale this cycle (!), but we don't refer to 3.1 as Yale's GPA floor because assuming you will be that one-in-a-thousand (literally) person, you would be making a foolhardy assumption.

I think a decent proxy is 5%, which is a probability which starts to fall outside the traditional definition of statistical definition. I looked at mylsn and went for the GPA at which the selected sample's acceptance rate first dropped to 5% for the past three years, non-ED, non-URM. I realize this is not an ideal metric, because it includes a sample consisting of some well below the stated "floor" (such that those who are right at it would have a greater chance than 5%), but I found the data to roughly conform to what people seem to mean when they use the term, which gives it value.

Yale: 3.83
Stanford: 3.83
Harvard: 3.62
Columbia: 3.30
Chicago: 3.50
NYU: 3.17
Penn: 3.16
Virginia: 3.17
Michigan: 3.07
Berkeley: 3.53
Duke: 3.36
Northwestern: n<30
Cornell: 3.25
Georgetown: n<30

This method would suggest, however, that schools who admit a limited number of students in a band where not many applicants are filed (think the 2.5-3.0 range for schools like NU and GULC) have a "low floor" when the real chances for that group may not be particularly high regardless of LSAT (obviously, it depends on the particular scenario). Those above however, look similar to the commonly quoted GPA floors bandied about on TLS.

Here is another method, which is derived by the tenth-lowest non-ED, non-URM GPA accepted the past three cycles, as per mylsn (note this bias toward a higher GPA floor for smaller schools). Ten is an arbitrarily selected number at which I personally felt an acceptance of a GPA can no longer be considered a "fluke", so to speak:

Yale: 3.83
Stanford: 3.77
Harvard: 3.62
Columbia: 3.40
Chicago: 3.52
NYU: 3.30
Penn: 3.35
Virginia: 3.29
Michigan: 3.25
Berkeley: 3.50
Duke: 3.40
Northwestern: 2.70
Cornell: 3.30
Georgetown: 3.08

Different numbers-based variations can conclude a different "GPA floor" for each school, so pick your method.

[jbagelboy]

Attempting to establish a numerical value on "GPA floor" requires a more precise definition of the term. Rarely do we use the phrase to literally mean "the lowest GPA at which a school will accept an applicant." For example, a non-URM with a 3.1 was accepted to Yale this cycle (!), but we don't refer to 3.1 as Yale's GPA floor because assuming you will be that one-in-a-thousand (literally) person, you would be making a foolhardy assumption.

I think a decent proxy is 5%, which is a probability which starts to fall outside the traditional definition of statistical definition. I looked at mylsn and went for the GPA at which the selected sample's acceptance rate first dropped to 5% for the past three years, non-ED, non-URM. I realize this is not an ideal metric, because it includes a sample consisting of some well below the stated "floor" (such that those who are right at it would have a greater chance than 5%), but I found the data to roughly conform to what people seem to mean when they use the term, which gives it value.

Yale: 3.83
Stanford: 3.83
Harvard: 3.62
Columbia: 3.30
Chicago: 3.50
NYU: 3.17
Penn: 3.16
Virginia: 3.17
Michigan: 3.07
Berkeley: 3.53
Duke: 3.36
Northwestern: n<30
Cornell: 3.25
Georgetown: n<30

This method would suggest, however, that schools who admit a limited number of students in a band where not many applicants are filed (think the 2.5-3.0 range for schools like NU and GULC) have a "low floor" when the real chances for that group may not be particularly high regardless of LSAT (obviously, it depends on the particular scenario). Those above however, look similar to the commonly quoted GPA floors bandied about on TLS.

Here is another method, which is derived by the tenth-lowest non-ED, non-URM GPA accepted the past three cycles, as per mylsn (note this bias toward a higher GPA floor for smaller schools). Ten is an arbitrarily selected number at which I personally felt an acceptance of a GPA can no longer be considered a "fluke", so to speak:

Yale: 3.83
Stanford: 3.77
Harvard: 3.62
Columbia: 3.40
Chicago: 3.52
NYU: 3.30
Penn: 3.35
Virginia: 3.29
Michigan: 3.25
Berkeley: 3.50
Duke: 3.40
Northwestern: 2.70
Cornell: 3.30
Georgetown: 3.08

Different numbers-based variations can conclude a different "GPA floor" for each school, so pick your method.

Nice

[nonexpostfacto]

Attempting to establish a numerical value on "GPA floor" requires a more precise definition of the term. Rarely do we use the phrase to literally mean "the lowest GPA at which a school will accept an applicant." For example, a non-URM with a 3.1 was accepted to Yale this cycle (!), but we don't refer to 3.1 as Yale's GPA floor because assuming you will be that one-in-a-thousand (literally) person, you would be making a foolhardy assumption.

I think a decent proxy is 5%, which is a probability which starts to fall outside the traditional definition of statistical definition. I looked at mylsn and went for the GPA at which the selected sample's acceptance rate first dropped to 5% for the past three years, non-ED, non-URM. I realize this is not an ideal metric, because it includes a sample consisting of some well below the stated "floor" (such that those who are right at it would have a greater chance than 5%), but I found the data to roughly conform to what people seem to mean when they use the term, which gives it value.

Yale: 3.83
Stanford: 3.83
Harvard: 3.62
Columbia: 3.30
Chicago: 3.50
NYU: 3.17
Penn: 3.16
Virginia: 3.17
Michigan: 3.07
Berkeley: 3.53
Duke: 3.36
Northwestern: n<30
Cornell: 3.25
Georgetown: n<30

This method would suggest, however, that schools who admit a limited number of students in a band where not many applicants are filed (think the 2.5-3.0 range for schools like NU and GULC) have a "low floor" when the real chances for that group may not be particularly high regardless of LSAT (obviously, it depends on the particular scenario). Those above however, look similar to the commonly quoted GPA floors bandied about on TLS.

Here is another method, which is derived by the tenth-lowest non-ED, non-URM GPA accepted the past three cycles, as per mylsn (note this bias toward a higher GPA floor for smaller schools). Ten is an arbitrarily selected number at which I personally felt an acceptance of a GPA can no longer be considered a "fluke", so to speak:

Yale: 3.83
Stanford: 3.77
Harvard: 3.62
Columbia: 3.40
Chicago: 3.52
NYU: 3.30
Penn: 3.35
Virginia: 3.29
Michigan: 3.25
Berkeley: 3.50
Duke: 3.40
Northwestern: 2.70
Cornell: 3.30
Georgetown: 3.08

Different numbers-based variations can conclude a different "GPA floor" for each school, so pick your method.

Nice

Wow, cool analysis. Thank you so much.

Do you think that these numbers are going to fall in the next few years, or do you think that they will basically stay the same? Or is it even possible to predict? I showed this to a family friend who graduated law school (T20) six years ago, and he couldn't believe it... He was actually angry, because he thinks that he might have been admitted to a much better school in today's climate.

[ScottRiqui]

Has anyone looked at LSN's data and compared it to the schools' published number of acceptances/matriculants? I'm curious how well LSN represents the accepted and attending pools of applicants.

[Ti Malice]

For example, a non-URM with a 3.1 was accepted to Yale this cycle (!)

Seriously? Who is this person? He or she is the former president of which country? I'm not sure I'll believe this until it shows up in next year's viewbook. They haven't gone lower than 3.49 for any applicant in several years.

[LSATSCORES2012]

Has anyone looked at LSN's data and compared it to the schools' published number of acceptances/matriculants? I'm curious how well LSN represents the accepted and attending pools of applicants.

Regulus has done this for applicants/acceptances, but not matriculants.

[ScottRiqui]

Has anyone looked at LSN's data and compared it to the schools' published number of acceptances/matriculants? I'm curious how well LSN represents the accepted and attending pools of applicants.

Regulus has done this for applicants/acceptances, but not matriculants.

Thanks - that's a great resource. So it appears that fewer than 15% of the applications to the T14 are reflected on LSN, and those applicants who *are* on LSN are significantly stronger than the applicant pool as a whole.

[jbagelboy]

Wow, cool analysis. Thank you so much.

Do you think that these numbers are going to fall in the next few years, or do you think that they will basically stay the same? Or is it even possible to predict? I showed this to a family friend who graduated law school (T20) six years ago, and he couldn't believe it... He was actually angry, because he thinks that he might have been admitted to a much better school in today's climate.

Recall that these are GPA floors, not medians. Your family friend may be conflating the occurrence of acceptance to a school with a certain gpa with the generic ability of people with that GPA gain admission on a regular basis. Very, very few non-urm 3.65 applicants will get into Harvard, or Columbia with a 3.35, or Cornell with a 3.2. These applicants must have had unique mitigating factors (high LSAT, non-trad, ridiculous soft) that would have encouraged adcoms to look past their GPA in prior years as well. In fact, law school admission at top schools is notably harder than it was a decade ago, although definitely easier than the past 2-3 years.

Now, we can all agree that the emphasis on LSAT has only increased since your family friends' time, and applications have become arguably more splitter-friendly with more leniency on the GPA. Still, no one circa 2004 should look at this data and say, "oh my 3.2 would totally have gotten me into Penn if Id applied now".

[Monochromatic Oeuvre]

For example, a non-URM with a 3.1 was accepted to Yale this cycle (!)

Seriously? Who is this person? He or she is the former president of which country? I'm not sure I'll believe this until it shows up in next year's viewbook. They haven't gone lower than 3.49 for any applicant in several years.

This is the poster. He was a 3.1/177 as he reported it. As I understood it (and I can't remember where exactly I read this), he had some insane story for terrible first-year grades and a 3.8+ the rest of the way, plus baller softs.

[Monochromatic Oeuvre]

Wow, cool analysis. Thank you so much.

Do you think that these numbers are going to fall in the next few years, or do you think that they will basically stay the same? Or is it even possible to predict? I showed this to a family friend who graduated law school (T20) six years ago, and he couldn't believe it... He was actually angry, because he thinks that he might have been admitted to a much better school in today's climate.

Recall that these are GPA floors, not medians. Your family friend may be conflating the occurrence of acceptance to a school with a certain gpa with the generic ability of people with that GPA gain admission on a regular basis. Very, very few non-urm 3.65 applicants will get into Harvard, or Columbia with a 3.35, or Cornell with a 3.2. These applicants must have had unique mitigating factors (high LSAT, non-trad, ridiculous soft) that would have encouraged adcoms to look past their GPA in prior years as well. In fact, law school admission at top schools is notably harder than it was a decade ago, although definitely easier than the past 2-3 years.

Now, we can all agree that the emphasis on LSAT has only increased since your family friends' time, and applications have become arguably more splitter-friendly with more leniency on the GPA. Still, no one circa 2004 should look at this data and say, "oh my 3.2 would totally have gotten me into Penn if Id applied now".

GPA floors are probably the least elastic as a school doesn't report them to anyone and is theoretically unconcerned with what they might be. The only time they would change is if you saw an appreciable change in the overall change in the applicant pool. I believe they've loosened a little over the past few years (maybe by a tenth of a point or so? I don't know exactly) but due to the nature of the statistic you need a large change in applicant quality to affect even a small change in the GPA floor, assuming admissions procedure stays consistent.

It can change rapidly, however, if admissions procedure changes in response to, presumably, the much more visible median statistics. Northwestern, for example, very shamelessly accepted basically anyone with a 170 this cycle, even when students had absolutely awful GPAs. If you're just using numbers as qualifications, you'd never take a 2.75/170 over a 3.25/169, but when you game medians your incentives change. NU knows full well it could fill a class of 24.9% awful GPAs without any reporting consequences if it needed to secure its LSAT median.

[shifty_eyed]

FWIW, I was accepted to Duke off the waitlist with a 3.29 hours after I sent in a LOCI.

[Monochromatic Oeuvre]

Oh, and here's some fun data to show anyone who thinks law school is so easy to get into nowadays. These are medians from Fall 1993, the first year following the implementation of the 180 scale:

Yale--3.83/171
Harvard--3.81/169
Stanford--3.68/167
Columbia--3.60/168
Chicago--3.64/169
NYU--3.60/168
Penn--3.70/165
UVA--3.50/164
Michigan--3.63/167
Berkeley--3.70/166
Duke--3.62/169
Northwestern--3.46/163
Cornell--3.56/165
Georgetown--3.52/166

2010 was the hardest year ever to get into law school. It has gotten easier since then, but this cycle probably still rates as one of the six or seven most difficult.

ETA: Medians are pulled from graphs, so give or take a few basis points on the GPA medians.

[Ramius]

Oh, and here's some fun data to show anyone who thinks law school is so easy to get into nowadays. These are medians from Fall 1993, the first year following the implementation of the 180 scale:

Yale--3.83/171
Harvard--3.81/169
Stanford--3.68/167
Columbia--3.60/168
Chicago--3.64/169
NYU--3.60/168
Penn--3.70/165
UVA--3.50/164
Michigan--3.63/167
Berkeley--3.70/166
Duke--3.62/169
Northwestern--3.46/163
Cornell--3.56/165
Georgetown--3.52/166

2010 was the hardest year ever to get into law school. It has gotten easier since then, but this cycle probably still rates as one of the six or seven most difficult.

ETA: Medians are pulled from graphs, so give or take a few basis points on the GPA medians.

Wasn't it 2003 that the ABA stoppped averaging LSATs though? I would think that would skew the numbers for LSAT downward at least slightly. I have no proof for it, and I'd also very willingly bet that test preparation materials has improved recently, but there must be some statistical variation based on that fact.

[finnandjake2]

Oh, and here's some fun data to show anyone who thinks law school is so easy to get into nowadays. These are medians from Fall 1993, the first year following the implementation of the 180 scale:

Yale--3.83/171
Harvard--3.81/169
Stanford--3.68/167
Columbia--3.60/168
Chicago--3.64/169
NYU--3.60/168
Penn--3.70/165
UVA--3.50/164
Michigan--3.63/167
Berkeley--3.70/166
Duke--3.62/169
Northwestern--3.46/163
Cornell--3.56/165
Georgetown--3.52/166

2010 was the hardest year ever to get into law school. It has gotten easier since then, but this cycle probably still rates as one of the six or seven most difficult.

ETA: Medians are pulled from graphs, so give or take a few basis points on the GPA medians.

Wasn't it 2003 that the ABA stoppped averaging LSATs though? I would think that would skew the numbers for LSAT downward at least slightly. I have no proof for it, and I'd also very willingly bet that test preparation materials has improved recently, but there must be some statistical variation based on that fact.

Interesting point. I wonder how much this skews the numbers. I also wonder if people prepped more for their first take knowing that a retake would only be averaged.

[Humbert Humbert]

Oh, and here's some fun data to show anyone who thinks law school is so easy to get into nowadays. These are medians from Fall 1993, the first year following the implementation of the 180 scale:

Yale--3.83/171
Harvard--3.81/169
Stanford--3.68/167
Columbia--3.60/168
Chicago--3.64/169
NYU--3.60/168
Penn--3.70/165
UVA--3.50/164
Michigan--3.63/167
Berkeley--3.70/166
Duke--3.62/169
Northwestern--3.46/163
Cornell--3.56/165
Georgetown--3.52/166

2010 was the hardest year ever to get into law school. It has gotten easier since then, but this cycle probably still rates as one of the six or seven most difficult.

ETA: Medians are pulled from graphs, so give or take a few basis points on the GPA medians.

Did Duke used to be ranked higher in USNWR?

[Monochromatic Oeuvre]

Wasn't it 2003 that the ABA stoppped averaging LSATs though? I would think that would skew the numbers for LSAT downward at least slightly. I have no proof for it, and I'd also very willingly bet that test preparation materials has improved recently, but there must be some statistical variation based on that fact.

Interesting point. I wonder how much this skews the numbers. I also wonder if people prepped more for their first take knowing that a retake would only be averaged.

Summer 2006. Between c/o 2009 (old reporting system) and c/0 2010 (new system), average median LSAT at the T14 rose 0.43 points--not a big factor in comparison to the much larger rise that occurred naturally as a result of other factors. Median LSAT had long been rising at the T14 well before the new reporting policy.

Did Duke used to be ranked higher in USNWR?

Yep. They were a top-ten school every year in the '90s. In the year those medians were used (1994), Duke was ranked #7, which is the highest the school was ever ranked. It's a little unfortunate for them that they're the only T14 whose median LSAT hasn't increased in 20 years, but there was indeed a time when Duke had the same median LSAT as Harvard.

[sillymike]

this is a great thread. props to nonexpostfacto for starting it, and for lsatscores2012 for posting that link for the analysis done by regulus!

Not trying to be "that guy", but taking this post and regulus' analysis together leads me to a question.
What kind of margin of error are we talking about here, if we're basing our data on lsn and most schools have 10% total applicants/20% admittances? Especially in the super splitter range, which is what I think the OP was driving at, it only takes one or two data points to completely change things.
Then again, maybe one or two data points won't change anything, because I'm SO not a stats guy.

Mostly, as a super splitter myself, I'm just wondering when people say things like "locked out of Penn", they're really saying "10% to 20% of the data shows you're locked out, but there's a 1/5 chance, or maybe less adjusting for applicant caliber, that they've actually dipped below your GPA this cycle and we just didn't know about it".

I love this thread. I hope it's one of the ones that never dies!

[Tiago Splitter]

The drop in applicants is obviously good for everybody, but a decline in a school's median may not really help out splitters, especially super splitters, very much. If a school's medians were 3.7/170 and then they move down to 3.7/169, suddenly hundreds of applicants with GPAs higher than yours but below median now have a median LSAT. If the LSAT median drops down to 168 things get even worse.

I have no idea if this is how it works in practice but it's something to think about.

[sillymike]

yeah. I always go back and forth on that point. Everybody says that this is good for splitters, and for sure some people are having mad cycles, but it's true that it might not be as great as it seems.
On the one hand, it does really just allow more people to be at median.
One the other hand, I would hope that if a schools lsat median starts dropping, they might be willing to take some lower GPAs to compensate or even try to raise the median. Numbers-wise, if a school's medians are 3.7 and 169, and they want a 3.7/170, isn't a 2.0/170 worth more for the median than a 3.69/169, since they will both drop the GPA median by the same amount (or not at all) but the 169 will (possibly) contribute to the median staying at 169 whereas the 170 could be the data point that secures the 170 median?
I mean, it looks pretty bad to take a 2.0/171 so i doubt many of the schools with a 170 median would actually see it that way, but from a pure math standpoint, isn't this the case? Sorry if I'm totally off, I'm still figuring out this whole median business

Is there a point where schools will start to think in terms of pure numbers, or do they care too much about their 25/75ths? I had assumed they don't care a huge amount about those numbers, but what do I know?
nothing. that's what.