Reason for Ever-Increasing LSAT medians Forum

[ahnhub]

A question about the way LSAT medians at schools across the board have been increasing this decade.

E.g: Michigan's 25-75 for LSAT:
2003: 164-168
2006: 164-169
2009: 167-170
2011: 168-171
(And there was in all likelihood a spike from 2000-2003 as well, because that's when law school apps went up dramatically for the first time--apps were down throughout the late '90s)

I know there are several reasons for the inflation. I'm wondering what peoples' thoughts are on what the biggest reason is. I know one reason it ISN'T: people getting smarter. The score is scaled to percentages.

1) More people applying: Definitely a reason. Apps go up, the number of people with higher scores goes up. But this can't explain everything, because apps haven't been going up as steadily as the medians--after an explosion of apps in 2003, there was actually a decline for several years, and we didn't see another explosion until the downturn in 2008. But medians kept pushing up the whole time. Also, you have to consider just how dramatic a rise in applicants has to be to account for pushing a school's median from the 94th percentile to the 97th percentile--basically half the scores which were over the median eight years ago are now below the median. I'm not a statistician but wouldn't that take something like a doubling of applications? There are NOT twice as many applicants today as in 2003, not even close.

2) Pressure from US News rankings to take only the best LSATs you can get: Also probably a factor, although elite schools will swear up and down it isn't. The pressure hasn't necessarily been ratcheting up every year, either, so it can't explain the continuous rise.

3) ABA deciding to publish highest scores, not the average: This happened in late 2006-early 2007. This is also probably a very big factor, in several ways--schools like Michigan are taking 165/171 retakers without thinking about it, probably, while before they had to consider the average score very seriously. Also, the policy provides an incentive for almost anyone to retake. Half the advice on TLS seems to be for people with perfectly good LSAT scores like 167 or 168 to retake 'til they break 170.

I think trying to find an answer to this question is meaningful, because it gives you a better sense of what these numbers mean. It's obvious schools are basically willing to take the highest LSAT score now, other things being equal: a 160/170 will blow a 165 out of the water in an admission cycle. But most of the T-14 schools continue to claim that they take into account all of your LSATs, even if they give most weight to the higher score.

It also should give us a better idea of what the applicant pool is gonna look like for 2011-12. If the biggest reason for the inflation is #1, we should be expecting an easing of the medians, because it's possible the number of applicants may drop all the way to 2003 levels. But if the biggest reason is #3 (I suspect it is), then the numbers will probably stay pretty close to last year (hopefully the medians won't somehow go up magically--if that happened I would really be suspicious).

I would love to hear peoples' thoughts!

[Real Madrid]

Test prep is becoming more widespread, and with every additional test administered there is more material to study. Also, I think the internet has played a part as a larger number of people nowadays spend significant time online at sites like TLS (gathering information from other people about stuff like the LSAT) than they did in 2003.

And fingers crossed on application numbers dropping DRASTICALLY this cycle. I'll take any help I can get.

[Bildungsroman]

Smarter applicants as more information becomes available and stratification of job opportunities by law school quality increases. Also, schools are more rankings-obsessed than in the past. These would make up my guess.

[Yeshia90]

Test prep is becoming more widespread, and with every additional test administered there is more material to study. Also, I think the internet has played a part as a larger number of people nowadays spend significant time online at sites like TLS (gathering information from other people about stuff like the LSAT) than they did in 2003.

And fingers crossed on application numbers dropping DRASTICALLY this cycle. I'll take any help I can get.

But the grading is curved. It doesn't matter if everyone does better, there are still going to be roughly the same percentages of 180s, 179s, 175s, etc.

The only answers are a) more applicants in the 165+ range (because of more applicants overall) and b) a greater reliance on LSAT scores in admissions.

Easy.

[Samara]

These are good answers, but I think #3 is the main reason. Look at your stats for Michigan: the increase from 2003 to 2006 was very small, possibly due to an increase in applications, but it's also within the bounds of statistical variation. The big jump happened from 2006 to 2009, the time period covering the change in retake policy. The increase from 2009 to 2011 is likely attributable to more people learning about the new retake policy and people retaking from before the policy change.

That said, I do think medians will come down, though probably not much. If you look at that chart of applicants by highest LSAT score that floats around here, you see a pretty significant drop across the board, with a slightly larger decrease, proportionally, in the 175+ range. T14 and TTT medians will likely drop, but I think T1/T2 medians will stay roughly the same as people will likely be more debt-averse this cycle and take a lower-ranked school for the scholly.

TL;DR - #3 is correct. Good news for T14 splitters.

[shastaca]

part of it is the higher number of applicants means higher schools can choose more selectively.

Michigan however has also discovered ways to gain the system by having a portion of their class not take the LSAT...which means smaller incoming class through admissions.

[Samara]

But the grading is curved. It doesn't matter if everyone does better, there are still going to be roughly the same percentages of 180s, 179s, 175s, etc.

It does matter because of the retake policy. Hypo: Jane, with prep, has a 75% of scoring a 175 if given three test administrations, but only a 40% chance if given one test administration. Under the old policy, Jane is likely to only take the test once and is thus less likely to score 175. Under the new policy, Jane is likely performing more prep since three tests is at least four months of extra prep, taking the test three times and is much more likely to score a 175.

Under the old rules, Jane is likely reported as a 168. Under the new rules, Jane is likely reported as a 175. Thus, though the score distributions are the same across each test, the distributions among applicants is skewed higher.

[Yeshia90]

But the grading is curved. It doesn't matter if everyone does better, there are still going to be roughly the same percentages of 180s, 179s, 175s, etc.

It does matter because of the retake policy. Hypo: Jane, with prep, has a 75% of scoring a 175 if given three test administrations, but only a 40% chance if given one test administration. Under the old policy, Jane is likely to only take the test once and is thus less likely to score 175. Under the new policy, Jane is likely performing more prep since three tests is at least four months of extra prep, taking the test three times and is much more likely to score a 175.

Under the old rules, Jane is likely reported as a 168. Under the new rules, Jane is likely reported as a 175. Thus, though the score distributions are the same across each test, the distributions among applicants is skewed higher.

I'd be very interested in seeing the % of applicants over time that were 1, 2, or 3-time test takers. That's a very interesting theory.

[JCougar]

But the grading is curved. It doesn't matter if everyone does better, there are still going to be roughly the same percentages of 180s, 179s, 175s, etc.

It does matter because of the retake policy. Hypo: Jane, with prep, has a 75% of scoring a 175 if given three test administrations, but only a 40% chance if given one test administration. Under the old policy, Jane is likely to only take the test once and is thus less likely to score 175. Under the new policy, Jane is likely performing more prep since three tests is at least four months of extra prep, taking the test three times and is much more likely to score a 175.

Under the old rules, Jane is likely reported as a 168. Under the new rules, Jane is likely reported as a 175. Thus, though the score distributions are the same across each test, the distributions among applicants is skewed higher.

This.

[scammedhard]

I'd be very interested in seeing the % of applicants over time that were 1, 2, or 3-time test takers. That's a very interesting theory.

Enjoy the reading:

"The Performance of Repeat Test Takers on the Law School Admission Test: 2003–2004 Through 2009–2010 Testing Years (TR 11-01)"
http://lsac.org/LsacResources/Research/TR/TR-11-01.pdf

[Samara]

I'd be very interested in seeing the % of applicants over time that were 1, 2, or 3-time test takers. That's a very interesting theory.

Enjoy the reading:

"The Performance of Repeat Test Takers on the Law School Admission Test: 2003–2004 Through 2009–2010 Testing Years (TR 11-01)"
http://lsac.org/LsacResources/Research/TR/TR-11-01.pdf

Great find! I just glanced over the report, but it looks to back up #3 as a driving force behind median increases. The charts on pages 5 and 6 show strong growth since 2007 in second-time test takers, both in raw number and proportionally. These trends will are likely to continue to increase for a couple years then plateau, no?

[09042014]

The LSAT isn't curve, it is equated. The percentage of high scores is increasing. Fairly dramatically in fact.

[scammedhard]

The LSAT isn't curve, it is equated. The percentage of high scores is increasing. Fairly dramatically in fact.

Figure 2 says otherwise; distributions of scores have been remarkably stable and consistent:

"LSAT Performance with Regional, Gender, and Racial/Ethnic Breakdowns: 2003–2004 Through 2009–2010 Testing Years (TR 10-03)"
http://www.lsac.org/LsacResources/Resea ... -10-03.pdf

[AffordablePrep]

One thing to remember is that many more people started taking the test during the economic collapse so the raw number of 170+ scores went up. As people generally try to get into the best ranked law school that they can, people in their early 20s generally have very similar "soft" credentials, and law school admissions make an effort to be objective, it only makes sense the medians will in fact go up. As the number of test takers goes back down to normal, it only makes sense that the medians will do likewise. However, as the TTT and TTTT realities are further exposed by the media, more people may end up retaking than would previously. The percentage of people that would bypass a year to retake the LSAT is very low as determined people in their early 20s feel a sense to "get onto the next stage of life" as quickly as possible without realizing sometimes the slow but steady route (retaking) might get them to their final destination much quicker as they don't get lost.

[KibblesAndVick]

The LSAT isn't curve, it is equated. The percentage of high scores is increasing. Fairly dramatically in fact.

Figure 2 says otherwise; distributions of scores have been remarkably stable and consistent:

"LSAT Performance with Regional, Gender, and Racial/Ethnic Breakdowns: 2003–2004 Through 2009–2010 Testing Years (TR 10-03)"
http://www.lsac.org/LsacResources/Resea ... -10-03.pdf

"Figure 2 shows the smoothed percentages of LSAT scores for all test takers (except
those who tested under nonstandard conditions) for each testing year from 2003–2004
through 2009–2010 (average LSAT scores were used for test takers who took the test
more than once in a given testing year). To control for the changing volume across
testing years, the number of test takers at each score was converted to a percentage
within each testing year. The smoothed percentages are very similar across testing
years."

1) The figure uses average test scores. So someone with a 165 and a 175retake is shown as a 170 in the chart. They're reported as a 175 when calculating a schools average LSAT score.

2) They're using the percentage of test takers at a given score in the chart. It's not showing the raw number of test takers who get a given score. It's showing the percentage of test takers who get a given score. Even if 1% of test takers get a 170 in both 2003 and 2009 there will still be more high end scores because the total number of test takers has increased.

[JamMasterJ]

#3

[ahduth]

#3

Yeah, pretty much. It should cause a unilateral increase in reported medians, which is what you're claiming.

[apollo2015]

Test prep is becoming more widespread, and with every additional test administered there is more material to study. Also, I think the internet has played a part as a larger number of people nowadays spend significant time online at sites like TLS (gathering information from other people about stuff like the LSAT) than they did in 2003.

And fingers crossed on application numbers dropping DRASTICALLY this cycle. I'll take any help I can get.

But the grading is curved. It doesn't matter if everyone does better, there are still going to be roughly the same percentages of 180s, 179s, 175s, etc.

The only answers are a) more applicants in the 165+ range (because of more applicants overall) and b) a greater reliance on LSAT scores in admissions.

Easy.

Perhaps the people who are most likely to be admitted to the higher ranked (USN&R) schools have increased their rate of test prep at a higher rate than those admitted to lower ranked schools.

If students going to TTT/TTTT schools have a lower level of information about law schools in general, it would be reasonable to suppose that they also know less about test prepping (or have less access to test prep resources).

Because the grading is curved the low-information students will now appear to have less ability than the high-information students.

[Samara]

Test prep is becoming more widespread, and with every additional test administered there is more material to study. Also, I think the internet has played a part as a larger number of people nowadays spend significant time online at sites like TLS (gathering information from other people about stuff like the LSAT) than they did in 2003.

And fingers crossed on application numbers dropping DRASTICALLY this cycle. I'll take any help I can get.

But the grading is curved. It doesn't matter if everyone does better, there are still going to be roughly the same percentages of 180s, 179s, 175s, etc.

The only answers are a) more applicants in the 165+ range (because of more applicants overall) and b) a greater reliance on LSAT scores in admissions.

Easy.

Perhaps the people who are most likely to be admitted to the higher ranked (USN&R) schools have increased their rate of test prep at a higher rate than those admitted to lower ranked schools.

If students going to TTT/TTTT schools have a lower level of information about law schools in general, it would be reasonable to suppose that they also know less about test prepping (or have less access to test prep resources).

Because the grading is curved the low-information students will now appear to have less ability than the high-information students.

I don't think there has been an appreciable increase in the availability of test prep and test prep information over the past 5-10 years though. Additionally, if raw test performance increased in the short term, the "curve" used to generate the scaled score would just become smaller, negating such an increase.