Does LSN skew to the left? Does that render it useless? Forum

[DeSilentio2728]

I have spent some time recently staring at the little yellow, green, and red dots on the graphs over at LSN and from my perspective they do not seem very representative of the median numbers from prior years, as they have a tendency to skew to the right for just about every law school.

For instance, if you look at Vanderbilt's LSAT numbers for last year:

Their 25th, Median, and 75th numbers were 164, 166, and 167 respectively.

For the 2008-2009 application cycle there were 236 admits and WL-admits reported on LSN for Vanderbilt during last year's cycle. Of these 236 admits: only 9 had scores at or below 164 (25th), only 31 had scores at or below 166 (median), and only 51 had scores at or below 167 (75th). This leaves a MASSIVE 185 applicants with scores above Vanderbilt's reported 75th. And if you look at the other schools (T-14, T-30, T-50, T-100) they all seem to be just as skewed to the right as Vanderbilt.

So my question being, is LSN just an esoteric device, and thus must it be rendered essentially useless?

[juevonate]

LSN isn't really esoteric.......just self-selecting the best, most dedicated applicants (and perhaps a few liars). Despite the fact that we might not see as many lower scores admitted, it still gives a pretty good idea of whether you have a shot or not. It has been the most accurate tool so far for predicting my acceptances, waitlists and denials.

Another thing to keep in mind is that the percentiles reflect the class, not just those accepted, so while, for example, Vanderbilt's class might skew to the "right" currently, after accepting the waitlisters, it may be more representative. It's likely that those accepted in June, July, and August are not as active in updating their statuses.

When I first read your post, I thought you were saying lsn was too conservative. I never got that impression....

[voice of reason]

LSN data are all self-reported and therefore, as you have discovered, not necessarily representative.

I think what it's useful for is telling us something about how numbers-driven the schools are. Some schools have lots of rejects mixed in with acceptances at the high end and have more acceptances mixed in with the rejects at the low end. Others have a sharper dividing line between the admission zone in the upper right and the rejection zone in the lower left. The former schools are looking at the whole applicant and the latter schools are playing it by the numbers. This may help tell you something about where it's worth sending a "reach" app.

But LSN should be taken with a grain of salt because the people posting numbers there are not representative of applicants as a whole.

[taw856]

Another thing to keep in mind is that the percentiles reflect the class, not just those accepted, so ... after accepting the waitlisters, it may be more representative.

This.

While it's true that people fudge numbers and better numbers self-select their way onto forums, to the degree it's possible to explain LSN's correlation with a class's eventual medians, it's this: Some yellow dots ultimately get in. Some green triangles ultimately attend elsewhere.

[DeSilentio2728]

LSN isn't really esoteric.......just self-selecting the best, most dedicated applicants (and perhaps a few liars). Despite the fact that we might not see as many lower scores admitted, it still gives a pretty good idea of whether you have a shot or not. It has been the most accurate tool so far for predicting my acceptances, waitlists and denials.

Another thing to keep in mind is that the percentiles reflect the class, not just those accepted, so while, for example, Vanderbilt's class might skew to the "right" currently, after accepting the waitlisters, it may be more representative. It's likely that those accepted in June, July, and August are not as active in updating their statuses.

When I first read your post, I thought you were saying lsn was too conservative. I never got that impression....

Doesn't that make it esoteric?

Where's the rest of the class? We all know how hard it is to get off of these waitlists, and so if this was representative, then it would be saying that the majority of people in Vanderbilt's entering class was accepted via WL not outright acceptance. Something tells me that is not the case.

[existenz]

Where's the rest of the class? We all know how hard it is to get off of these waitlists, and so if this was representative, then it would be saying that the majority of people in Vanderbilt's entering class was accepted via WL not outright acceptance. Something tells me that it is not correct.

First of all, people with better numbers are more likely to post on LSN. You know that in the real world only 10% of LSAT takers score 164 or higher. On LSN for this cycle alone, 50% of the people with profiles have scores of 164 or higher.

That said, the above responses gave you the correct answer. LSN tells you who was accepted. The medians you are analyzing involve who actually matriculated. For any school you look at (even Yale), the median LSAT/GPA of admits is going to be higher than the median of those who actually attend. Why? Because the people for whom a school is a reach are much more likely to attend compared to those with higher numbers who may have better options elsewhere.

[Borhas]

LSN isn't really esoteric.......just self-selecting the best, most dedicated applicants (and perhaps a few liars). Despite the fact that we might not see as many lower scores admitted, it still gives a pretty good idea of whether you have a shot or not. It has been the most accurate tool so far for predicting my acceptances, waitlists and denials.

Another thing to keep in mind is that the percentiles reflect the class, not just those accepted, so while, for example, Vanderbilt's class might skew to the "right" currently, after accepting the waitlisters, it may be more representative. It's likely that those accepted in June, July, and August are not as active in updating their statuses.

When I first read your post, I thought you were saying lsn was too conservative. I never got that impression....

Doesn't that make it esoteric?

Where's the rest of the class? We all know how hard it is to get off of these waitlists, and so if this was representative, then it would be saying that the majority of people in Vanderbilt's entering class was accepted via WL not outright acceptance. Something tells me that it is not correct.

some schools have gotten so many more admits that they may instead choose to get many from waitlists than to admit too many and risk overloading their classes

my theory anyway. I mean, isn't that the reasonable thing to do when your application go up by 20% or more two years in a row

[Genericswingman]

I would say that LSN probably represents somewhere between 10% and 20% of the incoming class, is skewed towards the higher-end of applicants and may be less meaningful as the application cycle continues. Most law schools are seeing application increases well above 20% (I've heard of two regional schools having year-over-year increases above 70%) which is probably better explained by the premise that applicants are applying to more schools than by the economy (I find it hard to believe that a bad economy would produce a 70% increase).

So, if my premise is correct, then there should be quite a bit of shuffling through the late spring and summer and a large number of applicants who defer enrollment until later years which should bring the distributions more in line with previous years.

Just a theory..

[djgoldbe]

I think, like someone already said, that a lot of those green dots on the higher end are going to other schools, meaning the actual admitted 'class' is far more towards the middle - especially with WL admits. However, if you look at Vanderbilt's graph for this year you see a clear line between green/yellow at the 167/3.7 mark. I don't think you would see such an obvious line in the sand if the numbers weren't representative. Some schools don't have such clear lines but a lot (like, say GW) also do.

[scribelaw]

I don't see how any of this is relevant to what LSN is actually useful for. You're making a good case that you can't look at LSN and see the makeup of an incoming class reflected. But you can still see how others with your numbers fared. Isn't that the point?

[dailygrind]

That's actually a skew to the left. The skew points in the direction of the tail.

[englawyer]

LSN can predict whether or not you will be admitted; it can't be used to estimate 25th, median, 75th etc. for the entire class.

[JCougar]

Where's the rest of the class? We all know how hard it is to get off of these waitlists, and so if this was representative, then it would be saying that the majority of people in Vanderbilt's entering class was accepted via WL not outright acceptance. Something tells me that it is not correct.

First of all, people with better numbers are more likely to post on LSN. You know that in the real world only 10% of LSAT takers score 164 or higher. On LSN for this cycle alone, 50% of the people with profiles have scores of 164 or higher.

That said, the above responses gave you the correct answer. LSN tells you who was accepted. The medians you are analyzing involve who actually matriculated. For any school you look at (even Yale), the median LSAT/GPA of admits is going to be higher than the median of those who actually attend. Why? Because the people for whom a school is a reach are much more likely to attend compared to those with higher numbers who may have better options elsewhere.

+1

[kazu]

That's actually a skew to the left. The skew points in the direction of the tail.

Haha I was wondering when someone would point this out.

I don't see how any of this is relevant to what LSN is actually useful for. You're making a good case that you can't look at LSN and see the makeup of an incoming class reflected. But you can still see how others with your numbers fared. Isn't that the point?

+1

[mountaintime]

for splitters like me, I don't think there is anything more useful for predicting admissions than LSN

[lawduder]

I found LSN to be extremely helpful in gauging what my chances were at certain schools. It's also nice in the sense that it gives you an idea of when decisions go out.

[DeSilentio2728]

Thanks for the responses. I just wanted to open up a discussion as to how useful LSN is to everyone.

Also, thanks for pointing out that the graph skews to the left. I am not as familiar with statistics as I should be.

[hotdog123]

for splitters like me, I don't think there is anything more useful for predicting admissions than LSN

This.

Seems like the LSAT cut-off lines are very obvious.

Extremely useful for the splitters.

[de5igual]

I have spent some time recently staring at the little yellow, green, and red dots on the graphs over at LSN and from my perspective they do not seem very representative of the median numbers from prior years, as they have a tendency to skew to the right for just about every law school.

For instance, if you look at Vanderbilt's LSAT numbers for last year:

Their 25th, Median, and 75th numbers were 164, 166, and 167 respectively.

For the 2008-2009 application cycle there were 236 admits and WL-admits reported on LSN for Vanderbilt during last year's cycle. Of these 236 admits: only 9 had scores at or below 164 (25th), only 31 had scores at or below 166 (median), and only 51 had scores at or below 167 (75th). This leaves a MASSIVE 185 applicants with scores above Vanderbilt's reported 75th. And if you look at the other schools (T-14, T-30, T-50, T-100) they all seem to be just as skewed to the right as Vanderbilt.

So my question being, is LSN just an esoteric device, and thus must it be rendered essentially useless?

Your lsat numbers are also off. For the 08-09 application cycle, Vandy had 164/168/169 as its 25/50/75 LSAT percentiles

[DeSilentio2728]

I have spent some time recently staring at the little yellow, green, and red dots on the graphs over at LSN and from my perspective they do not seem very representative of the median numbers from prior years, as they have a tendency to skew to the right for just about every law school.

For instance, if you look at Vanderbilt's LSAT numbers for last year:

Their 25th, Median, and 75th numbers were 164, 166, and 167 respectively.

For the 2008-2009 application cycle there were 236 admits and WL-admits reported on LSN for Vanderbilt during last year's cycle. Of these 236 admits: only 9 had scores at or below 164 (25th), only 31 had scores at or below 166 (median), and only 51 had scores at or below 167 (75th). This leaves a MASSIVE 185 applicants with scores above Vanderbilt's reported 75th. And if you look at the other schools (T-14, T-30, T-50, T-100) they all seem to be just as skewed to the right as Vanderbilt.

So my question being, is LSN just an esoteric device, and thus must it be rendered essentially useless?

Your lsat numbers are also off. For the 08-09 application cycle, Vandy had 164/168/169 as its 25/50/75 LSAT percentiles

I used the profile from LSN as my basis for the numbers. They stated that the information was based off of the class of 2009.

Updated numbers to reflect Vandy's reporting via their own website:

at or below 164 = 9 applicants

at or below 168 = 52 applicants

at or below 169 = 86 applicants

above 169 = 89 applicants

Still skews to the left but this can more easily (but not entirely) be explained now by the theory that higher qualified applicants choose to go to higher ranked schools even though they were admitted.

[de5igual]

I have spent some time recently staring at the little yellow, green, and red dots on the graphs over at LSN and from my perspective they do not seem very representative of the median numbers from prior years, as they have a tendency to skew to the right for just about every law school.

For instance, if you look at Vanderbilt's LSAT numbers for last year:

Their 25th, Median, and 75th numbers were 164, 166, and 167 respectively.

For the 2008-2009 application cycle there were 236 admits and WL-admits reported on LSN for Vanderbilt during last year's cycle. Of these 236 admits: only 9 had scores at or below 164 (25th), only 31 had scores at or below 166 (median), and only 51 had scores at or below 167 (75th). This leaves a MASSIVE 185 applicants with scores above Vanderbilt's reported 75th. And if you look at the other schools (T-14, T-30, T-50, T-100) they all seem to be just as skewed to the right as Vanderbilt.

So my question being, is LSN just an esoteric device, and thus must it be rendered essentially useless?

Your lsat numbers are also off. For the 08-09 application cycle, Vandy had 164/168/169 as its 25/50/75 LSAT percentiles

I used the profile from LSN as my basis for the numbers. They stated that the information was based off of the class of 2009.

c/o 2009 = 2005-2006 application cycle

[ruleser]

I would say that LSN probably represents somewhere between 10% and 20% of the incoming class, is skewed towards the higher-end of applicants and may be less meaningful as the application cycle continues. Most law schools are seeing application increases well above 20% (I've heard of two regional schools having year-over-year increases above 70%) which is probably better explained by the premise that applicants are applying to more schools than by the economy (I find it hard to believe that a bad economy would produce a 70% increase).

So, if my premise is correct, then there should be quite a bit of shuffling through the late spring and summer and a large number of applicants who defer enrollment until later years which should bring the distributions more in line with previous years.

Just a theory..

Expecting this- the 'defer enrollment to later years' might really show up since for the first time in a generation, people are debt adverse, just as tuitions have spiked. I expect a sig increase in 'cold feet'/plan b'ing

[pinkzeppelin]

I think another possible explanation is pride. People are proud when their strong numbers and softs get them into a good school. So when people do really well and get into their schools deservedly, they like to post their cycle on LSN to share that success with the world. People who are accepted into a law school with lower numbers might not be as proud. If your daddy donated one million to Vandy to get accepted, you probably wouldn't be too inclined to post your 155/3.2. Or if you golf with the dean of admissions and you got in with a 160/3.5, you likewise won't feel proud enough to want to share it.

So these are probably less common cases, but I think they are overwhelmingly saturated at the lower end of the applicant pool, causing a skew. The cycles of connected/super rich people aren't really applicable to the majority of us, so I'm okay with them not posting. Better not to give false hope.

[wired]

When I first started posting here, I was adamantly against LSN having any really meaningful information. However, I have changed my tune the more I have looked at the information. Most of the limitations everyone has discussed have been pretty straight on.

The place that LSN is most helpful is for traditional, white applicants who scored 160-175 on the LSAT and have over a 3.0 GPA. Everyone outside of that plays by different rules and doesn't have enough data points on LSN to make any really meaningful predictions. However, even the one's outside of that can make reference points from positive data on LSN. ("Another hispanic with a 165 3.7 got into Vanderbilt, I guess it is a possibility for me.")

For some hard data, look at this grid I made comparing GW's actual applicant pool and admissions versus LSN applicant pool and admissions. https://dl.dropbox.com/u/1748649/gwugrid.png

Within the 160-175, 3.0+ range, the biggest difference you see is an 11% difference for people who scored 165-169 and have a 3.0-3.25, where LSN understates the likelihood of that group getting in. That's obviously not good enough to be gospel, but it gives you a great understanding of your odds. If you look at just the median number applicants, I am sure it would be even more meaningful.