-15 curve almost guaranteed

iceland
Posts: 74
Joined: Tue Dec 14, 2010 11:21 am

Re: -15 curve almost guaranteed

Postby iceland » Tue Dec 28, 2010 3:31 pm

:!:
Last edited by iceland on Tue Jan 04, 2011 9:26 am, edited 1 time in total.

User avatar
robotclubmember
Posts: 743
Joined: Wed Sep 15, 2010 7:53 am

Re: -15 curve almost guaranteed

Postby robotclubmember » Tue Dec 28, 2010 3:35 pm

SchopenhauerFTW wrote:
robotclubmember wrote:
SchopenhauerFTW wrote:
robotclubmember wrote:Lol.

I think it's cute when people fail to recognize that adding one to the numerator of an expression and adding one to the denominator of an expression are actually two different things...


What an incredibly douchey and condescending/funny and lol-worthy comment.


Sorry brah.


No problem broseph. I simply thought to myself 'wow what a jerk comment, but how does this differ from how I react to the mistakes of others?'


Haha. I'd have reacted less harshly if this guy's introduction to the forum wasn't him spamming his vlogs: viewtopic.php?f=6&t=142210&p=3776640#p3776640

That and this whole thread is kinda clownish. I'm just tweaking like everyone else on the LSAT score, don't worry about it.

User avatar
KevinP
Posts: 1324
Joined: Sat Sep 26, 2009 8:56 pm

Re: -15 curve almost guaranteed

Postby KevinP » Tue Dec 28, 2010 3:37 pm

robotclubmember wrote:
KevinP wrote:
robotclubmember wrote:So because there was an extra question, you think that means you can get an extra 1% wrong and still get a 170? That's insane logic considering there has only been one time the LSAC has allowed a 170 with under 86% correct, and far more often has required test takers to get 90%+ for a 170.


Actually,
Feb '97, PTs 7, 9, 17, 23, 27 all gave the test taker the ability to score a 170 with under 86% correct.

Edit: Decided against sounding like a d-bag.


Oh shi-

OK. It still hasn't happened in over a decade and that was a different era of test-takers, before the days of TLS and established prep methods like PowerScore, etc.

In my defense, I wake and baked this morning.


Yeah, everyone makes mistakes. I doubt we'll see a -15 but I still think a -14 is possible. The LG section resembled something closer to the 20s and 30s in terms of difficulty and the established prep methods didn't really teach one how to tackle games 2 and 3 imo. I also found LR 1 to be difficult and I'm usually really good at estimating LR.

jblev1
Posts: 41
Joined: Mon Dec 27, 2010 9:55 am

Re: -15 curve almost guaranteed

Postby jblev1 » Wed Dec 29, 2010 11:08 pm

robotclubmember wrote:
jblev1 wrote:
robotclubmember wrote:
jblev1 wrote:With 102 questions? That means you have to get 89 correct for a 170! Or 88 with a -14! That is rough man, that low of a curve makes a high question test harder, not easier. I am going with OP on this one, -15 for real.


Lol.

I think it's cute when people fail to recognize that adding one to the numerator of an expression and adding one to the denominator of an expression are actually two different things.

In other words, adding one additional question to the number of questions you took does not mean that for things to be equal you should be allowed to get one extra question wrong. In fact, the opposite is probably truer.

Typical -14 Curve: 87/101 = 86.1% correct
Hypothetical PT62 -14 Curve: 88/102 = 86.2% correct
"Guaranteed" PT62 -15 Curve: 87/102 = 85.2% correct

So because there was an extra question, you think that means you can get an extra 1% wrong and still get a 170? That's insane logic considering there has only been one time the LSAC has allowed a 170 with under 86% correct, and far more often has required test takers to get 90%+ for a 170.

Lol, I'm done speculating on curve, it's almost here anyway. Some will argue "but we had less time," but I'd say the questions were slightly easier, and two 26Q LR sections is really not a big deal. Either way, the curve is determined by information we do not have access to, determined with metrics not fully known to us, so I lol at the conclusion that -15 is guaranteed because there was an extra question.


Yes, but the curve is determined by the percentile of scorers in the range for the past 3 years, not by the amount correct.


I made it clear that I'm not suggesting the amount correct determines the curve, all I did was point out a quick statistical analysis, not to support any conclusion of mine, but to undermine the soundness of this -15 junk. An extra question means a lot less than you seem to think. And what you said is not actually true either, it's determined through the experimental section performance. Actually, your sentence doesn't even mean anything. "percentile of scorers in the range for the past 3 years?" What range? What percentile? This is a nonsense statement.


Yes it is about the percentile of all scores for all tests in the past 3 years, I have heard. Not the scale of all scores, the percentile, meaning not the raw score, but the percentile in which you score compared to other scorers is how they decide the curve, and this is how they decide the scale to raw score as well.

User avatar
robotclubmember
Posts: 743
Joined: Wed Sep 15, 2010 7:53 am

Re: -15 curve almost guaranteed

Postby robotclubmember » Thu Dec 30, 2010 12:02 am

jblev1 wrote:
Yes it is about the percentile of all scores for all tests in the past 3 years, I have heard. Not the scale of all scores, the percentile, meaning not the raw score, but the percentile in which you score compared to other scorers is how they decide the curve, and this is how they decide the scale to raw score as well.


http://www.youtube.com/watch?v=nHj0dtSwxqQ

User avatar
ResolutePear
Posts: 8614
Joined: Fri Jul 02, 2010 10:07 pm

Re: -15 curve almost guaranteed

Postby ResolutePear » Thu Dec 30, 2010 12:09 am

Retake.

User avatar
well-hello-there
Posts: 320
Joined: Tue Oct 12, 2010 8:38 pm

Re: -15 curve almost guaranteed

Postby well-hello-there » Thu Dec 30, 2010 1:25 am

jblev1 wrote:
jblev1 wrote:Yes, but the curve is determined by the percentile of scorers in the range for the past 3 years, not by the amount correct.
Yes it is about the percentile of all scores for all tests in the past 3 years, I have heard. Not the scale of all scores, the percentile, meaning not the raw score, but the percentile in which you score compared to other scorers is how they decide the curve, and this is how they decide the scale to raw score as well.
All Wrong.
The curve on any given LSAT is determined by the performance of the test takers who had pieces of that test as their experimental sections.
The percentile that you see associated with your LSAT score is derived from the score distribution of all LSAT takers over the past 3 years.

jblev1
Posts: 41
Joined: Mon Dec 27, 2010 9:55 am

Re: -15 curve almost guaranteed

Postby jblev1 » Thu Dec 30, 2010 11:51 pm

well-hello-there wrote:
jblev1 wrote:
jblev1 wrote:Yes, but the curve is determined by the percentile of scorers in the range for the past 3 years, not by the amount correct.
Yes it is about the percentile of all scores for all tests in the past 3 years, I have heard. Not the scale of all scores, the percentile, meaning not the raw score, but the percentile in which you score compared to other scorers is how they decide the curve, and this is how they decide the scale to raw score as well.
All Wrong.
The curve on any given LSAT is determined by the performance of the test takers who had pieces of that test as their experimental sections.
The percentile that you see associated with your LSAT score is derived from the score distribution of all LSAT takers over the past 3 years.


No, the score is decided by the percentile.

User avatar
Bildungsroman
Posts: 5548
Joined: Sun Apr 11, 2010 2:42 pm

Re: -15 curve almost guaranteed

Postby Bildungsroman » Thu Dec 30, 2010 11:54 pm

jblev1 wrote:
well-hello-there wrote:
jblev1 wrote:
jblev1 wrote:Yes, but the curve is determined by the percentile of scorers in the range for the past 3 years, not by the amount correct.
Yes it is about the percentile of all scores for all tests in the past 3 years, I have heard. Not the scale of all scores, the percentile, meaning not the raw score, but the percentile in which you score compared to other scorers is how they decide the curve, and this is how they decide the scale to raw score as well.
All Wrong.
The curve on any given LSAT is determined by the performance of the test takers who had pieces of that test as their experimental sections.
The percentile that you see associated with your LSAT score is derived from the score distribution of all LSAT takers over the past 3 years.


No, the score is decided by the percentile.


Then why do the scores associated with each percentile change from administration to administration?

User avatar
well-hello-there
Posts: 320
Joined: Tue Oct 12, 2010 8:38 pm

Re: -15 curve almost guaranteed

Postby well-hello-there » Thu Dec 30, 2010 11:55 pm

jblev1 wrote:
well-hello-there wrote:
jblev1 wrote:
jblev1 wrote:Yes, but the curve is determined by the percentile of scorers in the range for the past 3 years, not by the amount correct.
Yes it is about the percentile of all scores for all tests in the past 3 years, I have heard. Not the scale of all scores, the percentile, meaning not the raw score, but the percentile in which you score compared to other scorers is how they decide the curve, and this is how they decide the scale to raw score as well.
All Wrong.
The curve on any given LSAT is determined by the performance of the test takers who had pieces of that test as their experimental sections.
The percentile that you see associated with your LSAT score is derived from the score distribution of all LSAT takers over the past 3 years.


No, the score is decided by the percentile.

The "score" is determined by the number of questions a test-taker answers correctly.

jblev1
Posts: 41
Joined: Mon Dec 27, 2010 9:55 am

Re: -15 curve almost guaranteed

Postby jblev1 » Fri Dec 31, 2010 12:05 am

The scale is the number that goes with the amount answered correctly, and the score is determined by the percentile for the past 3 years, they look at what percent scored in each raw score and then set a scale to that for LSAT scores.

That is how the curve is formulated, the percentile of correct answers by each test taker compared to the percentile of correct answers in previous years, and with this percentile they can see how they need to adjust the curve, and with the curve adjusted the scale is a gimme.

User avatar
well-hello-there
Posts: 320
Joined: Tue Oct 12, 2010 8:38 pm

Re: -15 curve almost guaranteed

Postby well-hello-there » Fri Dec 31, 2010 12:17 am

Bildungsroman wrote:
jblev1 wrote:No, the score is decided by the percentile.


Then why do the scores associated with each percentile change from administration to administration?

actually, the percentile associated with each scaled score changes slightly from administration to administration. For example, a 165 before the October test was good enough for the 92nd percentile. A 165 on the October test got you into the 92.1st percentile.
This is because relatively fewer people scored 165 or higher on the October test as compared with the average relative number of test takers over the past 3 years who scored 165 or higher. The October test by itself moving up the 165 scaled score percentile one tenth from 92 to 92.1 means that if the October test percentiles were calculated independent of the three year average, a 165 would be an even higher percentile. Say...93rd percentile maybe.
This is because the actual performance of all October test takers this year was worse than average at the part of the curve corresponding to a 165. I don't have the data points for each scaled score on the curve but you could do that kind of analysis for each score if you wanted to and had that information. Which I think is public info.

User avatar
well-hello-there
Posts: 320
Joined: Tue Oct 12, 2010 8:38 pm

Re: -15 curve almost guaranteed

Postby well-hello-there » Fri Dec 31, 2010 12:18 am

jblev1 wrote:The scale is the number that goes with the amount answered correctly, and the score is determined by the percentile for the past 3 years, they look at what percent scored in each raw score and then set a scale to that for LSAT scores.

That is how the curve is formulated, the percentile of correct answers by each test taker compared to the percentile of correct answers in previous years, and with this percentile they can see how they need to adjust the curve, and with the curve adjusted the scale is a gimme.

This is all wrong.

jblev1
Posts: 41
Joined: Mon Dec 27, 2010 9:55 am

Re: -15 curve almost guaranteed

Postby jblev1 » Fri Dec 31, 2010 12:22 am

well-hello-there wrote:
jblev1 wrote:The scale is the number that goes with the amount answered correctly, and the score is determined by the percentile for the past 3 years, they look at what percent scored in each raw score and then set a scale to that for LSAT scores.

That is how the curve is formulated, the percentile of correct answers by each test taker compared to the percentile of correct answers in previous years, and with this percentile they can see how they need to adjust the curve, and with the curve adjusted the scale is a gimme.

This is all wrong.


Dang, I just looked it up, you are right. That is how the SAT is scored, and there is no curve. I thought they did the same thing for the LSAT.

From wikipedia:

The LSAT system of scoring is predetermined and does not reflect test takers' percentile, unlike the SAT. The relationship between raw questions answered correctly (the "raw score") and scaled score is determined before the test is administered, through a process called equating.[12] This means that the conversion standard is set beforehand, and the distribution of percentiles can vary during the scoring of any particular LSAT.

User avatar
hokie
Posts: 339
Joined: Tue Jun 15, 2010 1:32 am

Re: -15 curve almost guaranteed

Postby hokie » Fri Dec 31, 2010 12:35 am

jblev1 wrote:
well-hello-there wrote:
jblev1 wrote:The scale is the number that goes with the amount answered correctly, and the score is determined by the percentile for the past 3 years, they look at what percent scored in each raw score and then set a scale to that for LSAT scores.

That is how the curve is formulated, the percentile of correct answers by each test taker compared to the percentile of correct answers in previous years, and with this percentile they can see how they need to adjust the curve, and with the curve adjusted the scale is a gimme.

This is all wrong.


Dang, I just looked it up, you are right. That is how the SAT is scored, and there is no curve. I thought they did the same thing for the LSAT.

From wikipedia:

The LSAT system of scoring is predetermined and does not reflect test takers' percentile, unlike the SAT. The relationship between raw questions answered correctly (the "raw score") and scaled score is determined before the test is administered, through a process called equating.[12] This means that the conversion standard is set beforehand, and the distribution of percentiles can vary during the scoring of any particular LSAT.


yeaaaaa..............I'm going to venture to say that this system doesn't help us [Dec LSAT people who got screwed on the LG) :roll: ........

User avatar
Pricer
Posts: 558
Joined: Tue Jan 19, 2010 9:25 pm

Re: -15 curve almost guaranteed

Postby Pricer » Fri Dec 31, 2010 1:37 am

hokie wrote:
jblev1 wrote:
well-hello-there wrote:
jblev1 wrote:The scale is the number that goes with the amount answered correctly, and the score is determined by the percentile for the past 3 years, they look at what percent scored in each raw score and then set a scale to that for LSAT scores.

That is how the curve is formulated, the percentile of correct answers by each test taker compared to the percentile of correct answers in previous years, and with this percentile they can see how they need to adjust the curve, and with the curve adjusted the scale is a gimme.

This is all wrong.


Dang, I just looked it up, you are right. That is how the SAT is scored, and there is no curve. I thought they did the same thing for the LSAT.

From wikipedia:

The LSAT system of scoring is predetermined and does not reflect test takers' percentile, unlike the SAT. The relationship between raw questions answered correctly (the "raw score") and scaled score is determined before the test is administered, through a process called equating.[12] This means that the conversion standard is set beforehand, and the distribution of percentiles can vary during the scoring of any particular LSAT.


yeaaaaa..............I'm going to venture to say that this system doesn't help us [Dec LSAT people who got screwed on the LG) :roll: ........


Unless it was determined beforehand that the LG was harder than normal, making the overall test harder than normal, increasing the curve to "equate" the score. I think.

User avatar
androstan
Posts: 2553
Joined: Mon Jul 19, 2010 8:07 am

Re: -15 curve almost guaranteed

Postby androstan » Mon Jan 03, 2011 12:35 pm

retake




Return to “LSAT Prep and Discussion Forum”

Who is online

Users browsing this forum: #gobroncos, BobBoblaw, dj9i27, goldenbear2020, nimbus cloud, SunDevil14, VMars and 30 guests