Falling LSAT #s = Future Applicants May Study Less Than You Forum

[LSAT Blog]

This article seems to suggest that LSAT medians will be dropping in the future due to fewer LSAT-takers and fewer law school applicants:

If the sharp drop in takers in June 2011 continues for the application year, or even if the downward trend in the percentage of takers actually applying to law school continues, law schools will be forced to choose between

-cutting first-year enrollment or

-dipping ever deeper into the applicant pool.

If law schools take the second course, and just fill seats, the result should be falling academic credentials of entering classes.

After all, as much as law schools want to keep their LSAT/UGPA high, it seems pretty clear that they also want to fill their seats, rather than turning down easy money.

So...will future LSAT-takers be able to get away with studying less than you do?

[Yeshia90]

It's equated, dude. Not as though there are 1000 180s given out every year. 170 is 98th percentile whether 10,000 people take the LSAT or if 20 people take it.

[phillipjg]

It's equated, dude. Not as though there are 1000 180s given out every year. 170 is 98th percentile whether 10,000 people take the LSAT or if 20 people take it.

I believe he's talking about the medians of the schools.

[ColtsFan88]

It's equated, dude. Not as though there are 1000 180s given out every year. 170 is 98th percentile whether 10,000 people take the LSAT or if 20 people take it.

Not completely true. 170 is usually 98th percentile because they are so good at equating it. But it doesn't have to be. The raw scores to scaled scores are equated but the perxentiles are not. That's why they change slightly with each test administration.

[ColumbiaChamp]

It's equated, dude. Not as though there are 1000 180s given out every year. 170 is 98th percentile whether 10,000 people take the LSAT or if 20 people take it.

Class size stays the same so they will have to dip into lower scores to fill the class. If 10,000 people wrote then only 10 180s vs. 1,000 people wrote and only 1 180 (assuming the test takers have become neither dumber nor smarter).

[KevinP]

I hope this is the case and the statistics the article presents is interesting, but I am worried about extrapolating. My two main issues are the number of October's test takers and the distribution of scores (maybe the decrease was among the people from the lower end of the score distributions).

It's equated, dude. Not as though there are 1000 180s given out every year. 170 is 98th percentile whether 10,000 people take the LSAT or if 20 people take it.

You use that word but I do not think it means what you think it means. I think you're confusing the terms curved and equated.

(June 1998 - February 2001) Percentiles:
170 = 98.21%
160 = 83.07 %

(2007-2010) Percentiles:
170 = 97.40%
160 = 80.40%

Also, the differences between the test dates are significant.
June (Mean = 151.68, SD = 10.51). Approximate result: A 170 is the 95.9th percentile, ~4.1% of June test takers score a 170+.
December (Mean = 150.11, SD = 9.92). Approximate result: A 170 is the 97.8th percentile, ~2.2% of December test takers score a 170+.*

*I used data from 2009-2010 for calculations

[LSAT Blog]

+1 to both ColumbiaChamp and KevinP

As KevinP suggested, the decrease in test-takers may be disproportionately from those on the lower end of the score distribution, as fewer will apply on a whim, so a higher % will be those who are more serious about going to law school and about getting a good score.

However, it's safe to say that the decrease in test-takers is significant enough that many schools will have to dig at least a little bit deeper into their applicant pools, leading them to admit more applicants with lower numbers.

[Robespierre]

In reading the article, my eyebrows shot up at the first sentence, which contains the blatantly wrong statement that the number of applicants for Fall '10 was down ten percent; actually it was up slightly. But presumably that was just a typo and the author meant Fall '11. He gets back on track by making the excellent point that not only has the number of test-takers fallen for the last few LSAT administrations, but the percentage of test-takers who actually apply to LS has been falling steadily for many years. So it's a pretty good analysis.

I wouldn't get too excited yet about a huge drop in applicants for '11-'12, though, because:

1. The numbers can turn on a dime, so we need to see the number of October 2011 test-takers before drawing conclusions about this application year. Everyone assumes it will be down, as it was in June, but let's see the figures. The economic news has been so unremittingly bad lately that a lot of people may have taken the LSAT to panic-flee towards three more years of school.

2. I don't think the author gives enough attention to the fact that law schools can simply enroll fewer people, thus partially offsetting a drop in applicants. In the Medians thread, it's come out that a long list of schools enrolled smaller classes for '11 than '10. Yet the author treats the number of LS seats as relatively static.

[paul34]

Many law schools have expanded their capacity during the boom. I don't think suddenly deciding to just not use that capacity will be all too easy. They still have to pay the bills. I'm not sure if this is a good analogy, but I think it would be a bit like Walmart expanding, and then deciding to simply sit on the extra space instead of selling more stuff in there. I don't believe they would do that.

Now, there is a limitation with the above analogy in that in retail, there are situations where not selling certain products is better (financially) than selling it and having to take the loss. However, I don't believe there is a similar limitation with law school. If you stuff more students in, there isn't really a "loss" in that regard.

If the top law schools begin doing what is discussed in the OP, I think most will follow suit. They would probably be willing to drop their medians if "everyone else" was also doing the same (so that the relative rankings would not be skewed).

[NYCLSATTutor]

In reading the article, my eyebrows shot up at the first sentence, which contains the blatantly wrong statement that the number of applicants for Fall '10 was down ten percent; actually it was up slightly. But presumably that was just a typo and the author meant Fall '11. He gets back on track by making the excellent point that not only has the number of test-takers fallen for the last few LSAT administrations, but the percentage of test-takers who actually apply to LS has been falling steadily for many years. So it's a pretty good analysis.

I wouldn't get too excited yet about a huge drop in applicants for '11-'12, though, because:

1. The numbers can turn on a dime, so we need to see the number of October 2011 test-takers before drawing conclusions about this application year. Everyone assumes it will be down, as it was in June, but let's see the figures. The economic news has been so unremittingly bad lately that a lot of people may have taken the LSAT to panic-flee towards three more years of school.

2. I don't think the author gives enough attention to the fact that law schools can simply enroll fewer people, thus partially offsetting a drop in applicants. In the Medians thread, it's come out that a long list of schools enrolled smaller classes for '11 than '10. Yet the author treats the number of LS seats as relatively static.

Speaking to point #1...does anyone know when the number of October 2011 test-takers will be released to the public?

[Robespierre]

For the June exam, the number came out around July 20, so maybe we can extrapolate mid-late November?

[Jeffort]

There is an important set of factors the article does not mention or appear to have taken into account that certainly influenced the statistics as well as the authors interpretations of the LSAC data.

The analysis and graphs seem to rely on a major flawed assumption and possibly other flaws that are fatal to the statistical accuracy of what the graphs attempt to depict.

First graph: The black line is meant to represent the number of individuals that took the LSAT each year. The language in the article, for instance "annual numbers for LSAT takers", also appears to assume the LSAC Volume summary annual figure represents the total number of people that took the test each cycle. The assumption is incorrect, invalidates the factual accuracy of the graph and thereby pulls the rug out from under any interpretations/conclusions based on it since it misrepresents the true facts.

The LSATS ADMINISTERED total number per annual cycle published in the LSAC chart http://www.lsac.org/LSACResources/Data/ ... stered.asp is not a headcount of the total number of people that took the test each year. Rather, it is a count of the total number of test booklets handed to test takers once checked in at a test center whether or not a score is cancelled/not reported. Students that took the test twice in the same period are counted twice in the numbers, students that took it three times account for three of the total number, etc.

The fatal flaw of the report is that it apparently fails to take into account repeat test taker data and misunderstood what the administered tests data chart represents.
http://www.lsac.org/LSACResources/Data/ ... erData.pdf

With statistical analysis, if you put incomplete data and flawed assumptions in, then inaccurate/unreliable garbage/comparisons/correlations and conclusions come out.

The volume of people taking the LSAT does appear to be declining for now, but probably not as significantly as the article can lead you to believe. I doubt it will push down the medians and LSAT percentiles of tier 1 law schools, especially not those of T14 schools, but it will be interesting to see what happens with many of the schools lower in the rankings.

This is just a hunch and I have no data to back it up since there is no empirical source for it, but I think one of the causes for the record high test volumes during 2009-2010 and 2010-2011 testing years and significant recent drops was the various significant LSAC policy changes over the last few years. There was the roughly two year long era that ended shortly before the June 2011 LSAT when the point of no return deadline to avoid having a cancel, absent or crappy score on your CAS report if you weren't ready or something last minute interfered with your registered test day was almost three weeks before the administration date. Another likely contributing cause for higher volume is the ABA reporting and law school admissions policy changes about focusing on students highest score and LSs not having to report the average LSAT score of admitted students that took it multiple times.



Anyway, those are my thoughts for now about this topic.

[LSAT Blog]

Jeffort, your points regarding LSATs administered vs. LSAT-takers are all well and good, and, to a certain extent, they do weaken the article's evidence. However, the take-away from the article isn't *so* adversely affected by that distinction.

Rather, I'm concerned with what law schools will do about the significant declines in the number of applicants/applications to law school - something that doesn't appear to be impacted by any concern you've raised.

[suspicious android]

Rather, I'm concerned with what law schools will do about the significant declines in the number of applicants/applications to law school

That's a pretty awesome link.

[addy11]

With fewer people taking the test and applying, there doesn't need to be any concomitant drop in the top schools.

It could always be that a very high percentage of the 170+ and 3.7+ apply and attend, whereas a large proportion of the people who would have to pay sticker at a school ranked 20 or below decide to try their hand at something else.

Another scenario: even if the number of people with these numbers drops, it could just mean that the law schools weigh their priorities differently. The conventional wisdom is that it's largely a numbers game, but it's not entirely so. It certainly could approach that asymptote, however. I'm not saying a 175 with a 3.8 is going to beat out a Rhodes Scholar with a 170 and a 3.8 for a spot at Yale, but where the distinctions are less egregious, things that may matter now (whether you went to HYP for undergrad or Community U, whether you were a chemical engineering and philosophy double major or majored in basket weaving, etc.) they may become subordinate or irrelevant.

(As a person who is planning on his numbers getting him in places, I still think this would be a terrible outcome!)

[mattviphky]

fiw, wustl received less applications that the year before, and so cut the number of seats in this year's class 15%

[Robespierre]

Rather, I'm concerned with what law schools will do about the significant declines in the number of applicants/applications to law school - something that doesn't appear to be impacted by any concern you've raised.

Law schools will:

- recruit a slightly smaller class;
- accept fewer people on the strength of their soft factors;
- accept slightly fewer low-scoring URMs for diversity's sake;
- accept slightly fewer low-scoring in-staters (state schools only, obviously);
- lay out a bit more scholarship money to attract high-scorers; and
- put more effort into wooing high-scorers once accepted.

Some combination of these strategies will allow them to avoid major declines in medians ... as long as the drop in applicants isn't catastrophic (20% or more).