Top Law Schools

banned
http://www.top-law-schools.com/forums/v ... 6&t=119285

I don't think there was an "Oh fuck" part of a section, aka zones or !kung. I think many October test takers were fooled into thinking the curve would be more generous because of the difficulty of those two. However, the other LG's were easy and so were the other RC's on that exam.

That being said, I do think the average question on this exam was more difficult than the average question on the October or June exam. What that translates to, I'm not sure. I thought that LG's were average difficulty, LR1 was below average, LR2 was above average to difficult, and RC was above average to difficult. I left the exam thinking it was a -12.

I'd say that's a good breakdown. In Oct people underestimated how ridiculously easy LR was. IMO that was the easiest LR on any test I've ever come across and games (besides zones) was not tricky at all.

This test, overall, felt more dense, and even people good at LG in general have been saying they had to guess on a few.

I would say -12. It is december after all...

bgood2texas wrote:I don't think there was an "Oh fuck" part of a section, aka zones or !kung. I think many October test takers were fooled into thinking the curve would be more generous because of the difficulty of those two. However, the other LG's were easy and so were the other RC's on that exam.

That being said, I do think the average question on this exam was more difficult than the average question on the October or June exam. What that translates to, I'm not sure. I thought that LG's were average difficulty, LR1 was below average, LR2 was above average to difficult, and RC was above average to difficult. I left the exam thinking it was a -12.

Yeah, I would agree with the fact that there wasn't really anything that stood out as being the "Oh fuck". I didn't take the October LSAT, so I can't speak in particular to zones or !kung, but from what I have read, the Voicemails and Webmails was on the level of difficulty similar to that of zones. However alot of your ability in those games is the ability to quickly realize how to diagram and then start plugging in the rules and hit the questions. I think there were a lot of people who were probably able to do this, but its not without spending a good 10+ minutes on it which probably detracted from their ability to do the fourth question. Same goes for individuals who skipped the third to go onto the fourth. Personally, I think I figured out how to diagram the third fairly quickly but then applying the rules to the diagram took some time. I definitely had the feeling that not many people were able to diagram and make all the inferences needed on both of the games. Therefore, they had to have missed at least a few, if not more. Just some food for thought on those two games.

On a last note, the fourth game provided a rule that seemed like it offered a lot of help when applied to the given variables. However, that rule had to be applied to two other of the rules in order to actually effectively figure it out. I don't know how well I did this, but I think that the first question on this game offered a lot of help for the rest of the problem for those who figured it out.

Cheers.

*EDIT* Maybe I should include that I'm a long time lurker, first time poster.

What do you think this means for a 160?

I received a 158 on the October LSAT and I am shooting for a 164 on this exam. I really hope I at least broke the 160 bracket, I would be devastated if I fell back into the 150's again. I felt more confident leaving this exam and I only had to guess on a couple due to time constraints. On the October LSAT I ran out of time on one whole RC passage and a bunch of LR questions, plus the dreaded zones.

I definitely nailed the games, either -0 to -2.

The RC felt normal to me, so I probably missed anywhere from -7 to -10, or possibly better.

On the LR, I moved through pretty methodically and could assume my normal score of -4 to -6 on each section.

Given this thought process I will probably fall around the 80 questions right mark or possibly even a few notches higher. Any assumptions on what I would need to keep from falling below a 160 or achieve my 164 goal?

Thanks!

go get some alcohol and do some christmas shopping

Coveted wrote:*EDIT* Maybe I should add, long time lurker, first time post.

What do you think this means for a 160?

I received a 158 on the October LSAT and I am shooting for a 164 on this exam. I really hope I at least broke the 160 bracket, I would be devastated if I fell back into the 150's again. I felt more confident leaving this exam and I only had to guess on a couple due to time constraints. On the October LSAT I ran out of time on one whole RC passage and a bunch of LR questions, plus the dreaded zones.

I definitely nailed the games, either -1 or -2.

The RC felt normal to me, so I probably missed anywhere from -7 to -10, or possibly better.

On the LR, I moved through pretty methodically and could assume my normal score of -4 to -6 on each section.

Given this thought process I will probably fall around the 80 questions right mark or possibly even a few notches higher. Any assumptions on what I would need to keep from falling below a 160 or achieve my 164 goal?

Thanks!

a raw score of 80 usually ranges around 160-163 iirc

edited grammar

mg7 wrote:

Coveted wrote:*EDIT* Maybe I should add, long time lurker, first time post.

What do you think this means for a 160?

I received a 158 on the October LSAT and I am shooting for a 164 on this exam. I really hope I at least broke the 160 bracket, I would be devastated if I fell back into the 150's again. I felt more confident leaving this exam and I only had to guess on a couple due to time constraints. On the October LSAT I ran out of time on one whole RC passage and a bunch of LR questions, plus the dreaded zones.

I definitely nailed the games, either -1 or -2.

The RC felt normal to me, so I probably missed anywhere from -7 to -10, or possibly better.

On the LR, I moved through pretty methodically and could assume my normal score of -4 to -6 on each section.

Given this thought process I will probably fall around the 80 questions right mark or possibly even a few notches higher. Any assumptions on what I would need to keep from falling below a 160 or achieve my 164 goal?

Thanks!

a raw score of 80 usually ranges around 160-163 iirc

edited grammar

Does the curve have as strong of an effect on the middle of the pack as it does on the 170+?

*Edit*

I love your avatar lol.

megagnarley wrote:I'd say that's a good breakdown. In Oct people underestimated how ridiculously easy LR was. IMO that was the easiest LR on any test I've ever come across and games (besides zones) was not tricky at all.

This test, overall, felt more dense, and even people good at LG in general have been saying they had to guess on a few.

I would say -12. It is december after all...

Can't speak specifically about the Oct test because this was my first time taking it, but I would have to agree that compared to other tests I have taken this one was fairly dense. Personally, I felt that the Mexican Songs passage was extremely difficult to fully pick apart due to the author's language usage and then the connection between paragraphs. Very poorly written passage as far as clarity in my opinion but I was able to decipher it, but detracted from my ability to get to the last passage with enough time to do all the questions. The insect and plant passage as well was pretty difficult for me considering it took me a few minutes to correctly make the right association of what the second paragraph was trying to get at. Luckily the third passage had a topic I was familiar with and allowed me to complete quicker or I would have been completely destroyed on the last passage.

Coveted wrote:

mg7 wrote:

Coveted wrote:*EDIT* Maybe I should add, long time lurker, first time post.

What do you think this means for a 160?

I received a 158 on the October LSAT and I am shooting for a 164 on this exam. I really hope I at least broke the 160 bracket, I would be devastated if I fell back into the 150's again. I felt more confident leaving this exam and I only had to guess on a couple due to time constraints. On the October LSAT I ran out of time on one whole RC passage and a bunch of LR questions, plus the dreaded zones.

I definitely nailed the games, either -1 or -2.

The RC felt normal to me, so I probably missed anywhere from -7 to -10, or possibly better.

On the LR, I moved through pretty methodically and could assume my normal score of -4 to -6 on each section.

Given this thought process I will probably fall around the 80 questions right mark or possibly even a few notches higher. Any assumptions on what I would need to keep from falling below a 160 or achieve my 164 goal?

Thanks!

a raw score of 80 usually ranges around 160-163 iirc

edited grammar

Does the curve have as strong of an effect on the middle of the pack as it does on the 170+?

*Edit*

I love your avatar lol.

thanks lol. i don't believe it does, but some of the more renowned TLSers could probably answer that more accurately

justonemoregame wrote:go get some alcohol and do some christmas shopping

In the West we have a bunch of stoners, so its more likely that people will be puffing the blunt, not to say that a few drinks aren't involved though. As far as christmas shopping, not a fan, giving people shit that they will use for a month and then buy the new version a year later is a waste of money. I will just tell everyone, hey I was going to give you a christmas present but I had to pay over $150 f****** dollars for a test and then another $100 for some people to f****** sort through my other school information. So heres a high five! Ho ho ho

I read a thought on another thread recently that made a lot of sense and is worth expanding upon because - to me at least - it wasn't a naturally super obvious notion. Stay with me now...

The December LSAT testing population likely has, at the least, a slightly lower proportion of very strong test takers. This is because the vast majority of people don't pull a 170+ out of their ass without some significant preparation. And it stands to reason that if you're serious enough about getting a 170+ to do significant preparation, it's at least in part likely because you're very serious about going to a top-notch school. And from there we postulate that if you're very serious about getting into a top-notch school, you have done your research and realize it's better to apply early in the cycle. Applying early in the cycle means a June test date (unless you took a Dec or Feb test with the intention of waiting a cycle, but that doesn't appear to be very common). If not June, then definitely Oct, but Oct seems to be a very strong "re-take" month.

December, on the other hand, doesn't seem to be a test month of choice for the very strong test takers because they don't need it. They took in June and rocked it, or re-took in October to squeeze out the extra couple points they needed to reach 173-175 and rocked it. December is likely more common for people who took in Oct and needed a re-take (and needed the re-take because they aren't very strong test-takers) or who took in June, then in Oct and still sucked, and still needed Dec.

Accordingly, the overall slightly weaker performance on the December LSAT correlates with what, at least in recent years, is a bigger curve. The bigger curve, of course, would be required to ensure a consistency among the percentage of takers who get each score (i.e. ensuring a consistent bell curve within statisitcal acceptability).

Now, you may say, "C'mon HD, it can't really make that much of a difference." I'd argue you're right, it doesn't. It seems to make about a 2 to 3 question difference which in the grand scheme of 99-101 questions is...wait, hold on while I crunch the numbers...about 2-3% difference in raw score.

Am I saying this is definitively the reason and we can now rest easy because we've cracked the mystery of the December curve? Of course not. But I believe it could be a contributing factor. Another possible contributing factor is that LSAC simply often makes the December test harder. But, if they are striving to ensure a fairly consistent level of difficulty across all administrations, we then have to revert back to the strength of the test-takers to find the source of the bigger curve.

I guess overall what I'm saying is that it's December 2nd and I have nothing better to do than pontificate about this stuff until I get my score. And if you're still reading, you obviously don't, either.

mg7 wrote:
a raw score of 80 usually ranges around 160-163 iirc

edited grammar

I have most of the LSATs from the past half a decade in front of me, and have taken a few looks at their scales. I have to go back to the October 2008 exam to find a scale that put a raw score of 80 (in this case, out of 100 questions) at anything less than a 163. Both June and Oct 2008 have an 80 at a 161.

Going back prior to those dates, I can next find a test in which 80 =/= at least 163 in September 2006. Oct 2005, June 2006 and Sept 2006 all had an 80 at 162. Dec and Jun 05 had it at 161.

In every exam since Oct 08, a raw score of 80 has led to 163-165, no more and no less. I suspect that this exam will stay consistent with that.

HawgDriver wrote:I read a thought on another thread recently that made a lot of sense and is worth expanding upon because - to me at least - it wasn't a naturally super obvious notion. Stay with me now...

The December LSAT testing population likely has, at the least, a slightly lower proportion of very strong test takers. This is because the vast majority of people don't pull a 170+ out of their ass without some significant preparation. And it stands to reason that if you're serious enough about getting a 170+ to do significant preparation, it's at least in part likely because you're very serious about going to a top-notch school. And from there we postulate that if you're very serious about getting into a top-notch school, you have done your research and realize it's better to apply early in the cycle. Applying early in the cycle means a June test date (unless you took a Dec or Feb test with the intention of waiting a cycle, but that doesn't appear to be very common). If not June, then definitely Oct, but Oct seems to be a very strong "re-take" month.

December, on the other hand, doesn't seem to be a test month of choice for the very strong test takers because they don't need it. They took in June and rocked it, or re-took in October to squeeze out the extra couple points they needed to reach 173-175 and rocked it. December is likely more common for people who took in Oct and needed a re-take (and needed the re-take because they aren't very strong test-takers) or who took in June, then in Oct and still sucked, and still needed Dec.

Accordingly, the overall slightly weaker performance on the December LSAT correlates with what, at least in recent years, is a bigger curve. The bigger curve, of course, would be required to ensure a consistency among the percentage of takers who get each score (i.e. ensuring a consistent bell curve within statisitcal acceptability).

Now, you may say, "C'mon HD, it can't really make that much of a difference." I'd argue you're right, it doesn't. It seems to make about a 2 to 3 question difference which in the grand scheme of 99-101 questions is...wait, hold on while I crunch the numbers...about 2-3% difference in raw score.

Am I saying this is definitively the reason and we can now rest easy because we've cracked the mystery of the December curve? Of course not. But I believe it could be a contributing factor. Another possible contributing factor is that LSAC simply often makes the December test harder. But, if they are striving to ensure a fairly consistent level of difficulty across all administrations, we then have to revert back to the strength of the test-takers to find the source of the bigger curve.

I guess overall what I'm saying is that it's December 2nd and I have nothing better to do than pontificate about this stuff until I get my score. And if you're still reading, you obviously don't, either.

Curves are determined by equating the test to other tests, not from the distribution of test takers' scores.

mrizza wrote:[Curves are determined by equating the test to other tests, not from the distribution of test takers' scores.

But how do you equate one test to another test without analyzing the performance on that test by the people who took it? If they look at the December results and say, "Dang, we need x% of testers to score (###) in order to make this test consistent/equitable with other administrations, but we only have y% right now," doesn't it make sense that they would have to then adjust the curve accordingly? And I don't mean that giant swings occur; that would arguably invalidate the test. We're talking about very small movements of a few questions from one test to another, which is a prime indicator of the consistency of the tests. My argument was that the overall pool of December test takers is likely slightly weaker than June or October, which would influence the curve necessary to achieve the consistent percentages.

I admit, you're talking to a guy that for a living currently flies a gigantic gatling gun around that happens to have an airplane attached to it; not a statistics professor. So I'm all ears if I've just completely got this thing wrong on all levels.

HawgDriver wrote:

mrizza wrote:[Curves are determined by equating the test to other tests, not from the distribution of test takers' scores.

But how do you equate one test to another test without analyzing the performance on that test by the people who took it? If they look at the December results and say, "Dang, we need x% of testers to score (###) in order to make this test consistent/equitable with other administrations, but we only have y% right now," doesn't it make sense that they would have to then adjust the curve accordingly? And I don't mean that giant swings occur; that would arguably invalidate the test. We're talking about very small movements of a few questions from one test to another, which is a prime indicator of the consistency of the tests. My argument was that the overall pool of December test takers is likely slightly weaker than June or October, which would influence the curve necessary to achieve the consistent percentages.

I admit, you're talking to a guy that for a living currently flies a gigantic gatling gun around that happens to have an airplane attached to it; not a statistics professor. So I'm all ears if I've just completely got this thing wrong on all levels.

the "curve" is determined by how previous takers did on their experimental sections, which then became our real sections. the overall pool's strength doesn't come into play, because its already preset by those folks.

HawgDriver wrote:

mrizza wrote:[Curves are determined by equating the test to other tests, not from the distribution of test takers' scores.

But how do you equate one test to another test without analyzing the performance on that test by the people who took it? If they look at the December results and say, "Dang, we need x% of testers to score (###) in order to make this test consistent/equitable with other administrations, but we only have y% right now," doesn't it make sense that they would have to then adjust the curve accordingly? And I don't mean that giant swings occur; that would arguably invalidate the test. We're talking about very small movements of a few questions from one test to another, which is a prime indicator of the consistency of the tests. My argument was that the overall pool of December test takers is likely slightly weaker than June or October, which would influence the curve necessary to achieve the consistent percentages.

I admit, you're talking to a guy that for a living currently flies a gigantic gatling gun around that happens to have an airplane attached to it; not a statistics professor. So I'm all ears if I've just completely got this thing wrong on all levels.

Who knows how much / if they tinker with the equating after the test is done. But the entire point of having experimental sections is to see how people do on new questions, and then rate their difficulty. They then pull from this pool of new questions, with their calculated difficulty, and assign an overall difficulty to the new test; which gives them the curve. Its common wisdom on these forums that the curves are already set in stone before the test is administered because of the equating process, but who knows. There is no evidence actually supporting what you're saying, but it could be going on in the background.

ETA: I flew in U-28As for a living, so I'm not a stats professor either.

bitsy wrote:the "curve" is determined by how previous takers did on their experimental sections, which then became our real sections. the overall pool's strength doesn't come into play, because its already preset by those folks.

mrizza wrote:Who knows how much / if they tinker with the equating after the test is done. But the entire point of having experimental sections is to see how people do on new questions, and then rate their difficulty. They then pull from this pool of new questions, with their calculated difficulty, and assign an overall difficulty to the new test; which gives them the curve. Its common wisdom on these forums that the curves are already set in stone before the test is administered because of the equating process, but who knows. There is no evidence actually supporting what you're saying, but it could be going on in the background.

This makes sense to me. Thanks. Even though it seems strange to me that the relative strength of the population taking the actual test would have nothing to do with the curve (probably because I think of a test curve in the standard, traditional way), I should know better than to even waste time thinking about the processes that the gigantic brains at LSAC have in place. Like you said, it could could be going on in the background, but there's no evidence. I'll be on my way then...

mrizza wrote:ETA: I flew in U-28As for a living, so I'm not a stats professor either.

Nice.

The December LSAT testing population likely has, at the least, a slightly lower proportion of very strong test takers. This is because the vast majority of people don't pull a 170+ out of their ass without some significant preparation. And it stands to reason that if you're serious enough about getting a 170+ to do significant preparation, it's at least in part likely because you're very serious about going to a top-notch school. And from there we postulate that if you're very serious about getting into a top-notch school, you have done your research and realize it's better to apply early in the cycle. Applying early in the cycle means a June test date (unless you took a Dec or Feb test with the intention of waiting a cycle, but that doesn't appear to be very common). If not June, then definitely Oct, but Oct seems to be a very strong "re-take" month.

Of course, another theory is that the December test does include some strong testers since it is a relative no-brainer (since test averaging is out) for ~170's (Oct) trying to score some serious merit money with a mid-170.

Big Dog wrote:Of course, another theory is that the December test does include some strong testers since it is a relative no-brainer (since test averaging is out) for ~170's (Oct) trying to score some serious merit money with a mid-170.

Sure, I never said there weren't strong testers in December. I'm sure there are a lot of them. I just said it made sense to me that there would be fewer of them proportionally compared to June and October.

Websites and Voicemails game was like the Zones game?....Ha!
NOTHING is like the Zones game. The Zones game was everything that is wrong with the world. Websites and Voicemails game was doable. It was time consuming and I read one of the rules wrong but did catch it though and had enough time to finish the section. That didn't happen with Zones and I usually get -0 on games.

housebro13 wrote:Websites and Voicemails game was like the Zones game?....Ha!
NOTHING is like the Zones game. The Zones game was everything that is wrong with the world. Websites and Voicemails game was doable. It was time consuming and I read one of the rules wrong but did catch it though and had enough time to finish the section. That didn't happen with Zones and I usually get -0 on games.

When people say stuff like that they make me paranoid that I read a rule wrong and didn't catch it. I remember reading over the rules of the 2nd and 3rd game multiple times because it either seemed too easy or was wordy, damn this is going to eat at me for the rest of the year.

I know I am nearly a decade late with this, but, it couldn't be helped.

I am wondering if someone can help me figure something out. It has to do with the impact the of an outlier upon the perception of the relative difficulty of the December 2009 Logical Reasoning section. I mean someone who scored at or 1 off of 100 percent correct. Would this have the effect of creating a perception that the section was more difficult in general, when it could be actually that the test was about normal, but the curve was set by an outlier?

Is that a feasible hypothesis, or am I missing something?

Cheers

jshawncampbell wrote:I know I am nearly a decade late with this, but, it couldn't be helped.

I am wondering if someone can help me figure something out. It has to do with the impact the of an outlier upon the perception of the relative difficulty of the December 2009 Logical Reasoning section. I mean someone who scored at or 1 off of 100 percent correct. Would this have the effect of creating a perception that the section was more difficult in general, when it could be actually that the test was about normal, but the curve was set by an outlier?

Is that a feasible hypothesis, or am I missing something?

Cheers

Don't unnecessarily necro 5 year old posts please. Also the curve is set before the test if that answers your question