Top Law Schools

With the -14 curve on the Dec. test and this article to boot...wonder what this means for future tests...good/bad for us TLS'ers? Thoughts?

http://www.nytimes.com/2010/01/10/educa ... SAT&st=cse

....

I think the curve on the Dec. test was ridiculous. A certain score should not be so dependent on the number of people who are taking the test. I understand why they tried to normalize the score. But when tons of people take the test who normally wouldn't and many of whom shouldn't then it inflates the score of the top end. And I am sure that was the reason for the curve in Dec. Can you tell I didn't take it then?

By inflating the scores, the test makers assume a certain percentage of students will do well in law school, which makes little sense. I am sure this will result in a large number of unexpected rejections for those basing their predictions on last years numbers, especially at schools where the increase in application has been 30%+.

In the longer term this may actually shrink the curve. People will see the higher scores needed and study more and do better on the test. But because of the normalizer, more people doing better does not mean higher scores. The normalization of the number of test takers will mean there will be a normalization of the number of people with higher scores, even if they are doing better on the test. This may lead schools to fight over the students with more scholarship money to keep from sliding back to prerecession boom numbers. Honestly, I think Iowa and Indiana are playing this game already.

Of course, it is also possible the inflated scores could lead to a debunking or radical altering of USNWR rankings if the scores do not reflect the correlation of success in law school as much as it has in the past. It may also discourage the number of people who apply to law schools because they can't get the scores they see in the rankings. Which could be the answer to the over saturation of attorneys and excessive number of law schools in the US.

But this is all speculation.

[Uneducated curve argument that fails to understand statistics or how LSAC runs the show]

....

kittenmittons wrote:[Uneducated curve argument that fails to understand statistics or how LSAC runs the show]

Well can you educate us on the the curve? My understanding is that LSAC takes raw scores and turns them into the 120-180 score by using percentiles. Their reasoning is the varying degree of difficulty of the test. But the percentile normalization assumes equally qualified testing groups. I don't think that assumption holds up when there are an abnormal number of test takers because of the economy who seek law school as an alternative to a scary job market. As I said i think this group is less qualified as a whole, which makes the top look better by percentages than they would in a normal grouping.

Sauer Grapes wrote:MoS. The curve on any LSAT is said to be predetermined by using previous performance on experimental sections. How many people took the test, whether prepared or unprepared would not affect the curve.

If that is true then I am wrong. But why is a certain score consistently linked to a certain percentile if that is the case?

....

but i wanted to go into a fit of furious ITE rage

MoS wrote:

Sauer Grapes wrote:MoS. The curve on any LSAT is said to be predetermined by using previous performance on experimental sections. How many people took the test, whether prepared or unprepared would not affect the curve.

If that is true then I am wrong. But why is a certain score consistently linked to a certain percentile if that is the case?

It's not. 170 used to be 98.1 percentile, but as of Sept 09 it was the 97.5 percentile.

Unless you think that the surplus test-takers are somehow markedly different from the average test taker over the last five years, then the pre-established percentiles would fit the pool of surplus test-takers. An increase in the pool of applicants would mean an increase in the raw number of people scoring 170+. With a fixed demand for law students at the top-tier schools, the supply of 170+ (or pick your LSAT score) scorers will increase and schools will be able to raise their median LSAT score as a result. It's that simple, and it scares me.

Perhaps the marginal (in the sense that he/she would not have applied to law school if the recession had not happened) candidate is less strong in other areas and the pool of applicants that have high LSAT scores AND a compelling path to law school will not rise, therefore not affecting admissions too much. But then again I just use this hypothesis to ease my fears...

http://en.wikipedia.org/wiki/Equating

Read up.

You are basically accusing the LSAC of being completely incompetent when it is you who misunderstands their statistical processes.

All this curve talk misses the more important question: There's a NYT troll on TLS! But who? ...OP, I'm looking at you!

Woozy wrote:http://en.wikipedia.org/wiki/Equating

Read up.

You are basically accusing the LSAC of being completely incompetent when it is you who misunderstands their statistical processes.

Who were you responding to?

I was shocked to see a -14 curve in December.. its not often LSAC responds to my requests..

astro1819 wrote:

Woozy wrote:http://en.wikipedia.org/wiki/Equating

Read up.

You are basically accusing the LSAC of being completely incompetent when it is you who misunderstands their statistical processes.

Who were you responding to?

To MoS. Sorry for the confusion. I agree with you that there will be a larger number of high scorers as a result of the larger number of test takers, and schools may try to boost medians. Your second paragraph is unfortunately best described as wishful thinking, but I think you know that already.

Desert Fox wrote:

MoS wrote:

Sauer Grapes wrote:MoS. The curve on any LSAT is said to be predetermined by using previous performance on experimental sections. How many people took the test, whether prepared or unprepared would not affect the curve.

If that is true then I am wrong. But why is a certain score consistently linked to a certain percentile if that is the case?

It's not. 170 used to be 98.1 percentile, but as of Sept 09 it was the 97.5 percentile.

That would make me wrong.
But

IHaveDietMoxie wrote:but i wanted to go into a fit of furious ITE rage

I have yet to see any compelling evidence that the number of quality, competitive applications is up, even though this is like the 432,567 thread about the increase in the test takers.

MoS wrote:I think the curve on the Dec. test was ridiculous. A certain score should not be so dependent on the number of people who are taking the test. I understand why they tried to normalize the score. But when tons of people take the test who normally wouldn't and many of whom shouldn't then it inflates the score of the top end. And I am sure that was the reason for the curve in Dec. Can you tell I didn't take it then?

By inflating the scores, the test makers assume a certain percentage of students will do well in law school, which makes little sense. I am sure this will result in a large number of unexpected rejections for those basing their predictions on last years numbers, especially at schools where the increase in application has been 30%+.

In the longer term this may actually shrink the curve. People will see the higher scores needed and study more and do better on the test. But because of the normalizer, more people doing better does not mean higher scores. The normalization of the number of test takers will mean there will be a normalization of the number of people with higher scores, even if they are doing better on the test. This may lead schools to fight over the students with more scholarship money to keep from sliding back to prerecession boom numbers. Honestly, I think Iowa and Indiana are playing this game already.

Of course, it is also possible the inflated scores could lead to a debunking or radical altering of USNWR rankings if the scores do not reflect the correlation of success in law school as much as it has in the past. It may also discourage the number of people who apply to law schools because they can't get the scores they see in the rankings. Which could be the answer to the over saturation of attorneys and excessive number of law schools in the US.

But this is all speculation.

Nothing you said made any sense. I assume you're a social sciences major?

kalvano wrote:I have yet to see any compelling evidence that the number of quality, competitive applications is up, even though this is like the 432,567 thread about the increase in the test takers.

This guy is right. Some of the number make it seem like this cycle will be ultra competitive, but just look at LSN. That is the most relevant data (who got in, with what numbers) that you can get in real time. There, it seems everything is sort of going as predicted.

rayiner wrote:

MoS wrote:I think the curve on the Dec. test was ridiculous. A certain score should not be so dependent on the number of people who are taking the test. I understand why they tried to normalize the score. But when tons of people take the test who normally wouldn't and many of whom shouldn't then it inflates the score of the top end. And I am sure that was the reason for the curve in Dec. Can you tell I didn't take it then?

By inflating the scores, the test makers assume a certain percentage of students will do well in law school, which makes little sense. I am sure this will result in a large number of unexpected rejections for those basing their predictions on last years numbers, especially at schools where the increase in application has been 30%+.

In the longer term this may actually shrink the curve. People will see the higher scores needed and study more and do better on the test. But because of the normalizer, more people doing better does not mean higher scores. The normalization of the number of test takers will mean there will be a normalization of the number of people with higher scores, even if they are doing better on the test. This may lead schools to fight over the students with more scholarship money to keep from sliding back to prerecession boom numbers. Honestly, I think Iowa and Indiana are playing this game already.

Of course, it is also possible the inflated scores could lead to a debunking or radical altering of USNWR rankings if the scores do not reflect the correlation of success in law school as much as it has in the past. It may also discourage the number of people who apply to law schools because they can't get the scores they see in the rankings. Which could be the answer to the over saturation of attorneys and excessive number of law schools in the US.

But this is all speculation.

Nothing you said made any sense. I assume you're a social sciences major?

+1. Use of, and speculation based on, statistical analysis should require a license IMO.

Because I'm a nice guy, here's one of the better explanations available (link). If we can move this can of stupid (discussion about the curve) to that thread, it would do TLS a great favor.