LR Question Level Frequency Forum

[hiltopp01]

Does anybody know how many questions of each difficulty level there are on the LR section--going by Cambridge's difficulty scale? So, how many level 1 questions, level 2, level 3, level 4?

[KDLMaj]

Does anybody know how many questions of each difficulty level there are on the LR section--going by Cambridge's difficulty scale? So, how many level 1 questions, level 2, level 3, level 4?

It's different every section/test. That's why the scaled score:raw score conversion is different every test. Contrary to popular belief, it isn't curved around the performance of the test takers. (Mostly). It's determined through a process called pre-equating. They know the conversion before anyone has taken the exam.

[Adrian Monk]

does anyone know generally how easy diffcult is the jun/sep/dec/feb test? any patterns?

[KDLMaj]

does anyone know generally how easy diffcult is the jun/sep/dec/feb test? any patterns?

Feb tends to have a more forgiving curve for whatever reason.

But honestly- a more forgiving curve just means more challenging questions. It really does all even out in the end. Apply based on what's best for the admissions cycle and your prep process. Don't try to target a particular time of the year just because you think it may be easier/harder. The LSAT's pre-equating ALWAYS compensates for that.

[Jeffort]

does anyone know generally how easy diffcult is the jun/sep/dec/feb test? any patterns?

Feb tends to have a more forgiving curve for whatever reason.

But honestly- a more forgiving curve just means more challenging questions. It really does all even out in the end. Apply based on what's best for the admissions cycle and your prep process. Don't try to target a particular time of the year just because you think it may be easier/harder. The LSAT's pre-equating ALWAYS compensates for that.

I agree with the bottom part, but there is zero data to support your claim about feb scales. The most recent known feb scale is from the 2000 test, before LSAC started using the automated LSAT assembler software to produce test forms, so we have zero data about the scales of the last 14 Feb tests or any of them produced by the assembler program to current test specs. They could be super tight curves and easier tests for all we know, but more likely a fairly random mix where everything balances out anyway, making the whole issue really a moot point.

Take the test when you're ready to perform well and only then and don't worry about things outside your control, LSAC makes sure everything balances out so scaled scores properly represent the same skill/performance level no matter which test form it was achieved on.

[KDLMaj]

does anyone know generally how easy diffcult is the jun/sep/dec/feb test? any patterns?

Feb tends to have a more forgiving curve for whatever reason.

But honestly- a more forgiving curve just means more challenging questions. It really does all even out in the end. Apply based on what's best for the admissions cycle and your prep process. Don't try to target a particular time of the year just because you think it may be easier/harder. The LSAT's pre-equating ALWAYS compensates for that.

I agree with the bottom part, but there is zero data to support your claim about feb scales. The most recent known feb scale is from the 2000 test, before LSAC started using the automated LSAT assembler software to produce test forms, so we have zero data about the scales of the last 14 Feb tests or any of them produced by the assembler program to current test specs. They could be super tight curves and easier tests for all we know, but more likely a fairly random mix where everything balances out anyway, making the whole issue really a moot point.

Take the test when you're ready to perform well and only then and don't worry about things outside your control, LSAC makes sure everything balances out so scaled scores properly represent the same skill/performance level no matter which test form it was achieved on.

You are correct that released data for Feb stopped at 2000. But there are benefits to being a massive prep company with thousands of student-reported data sets. From what I gathered while I worked for one (disclaimer- I was never someone who had direct access to that data- so I only heard it second-hand), this was true of pre-2000, and student-reported data showed that it continued to be true afterwards. But it was far from perfect data, so take it with a grain of salt.

[Jeffort]

does anyone know generally how easy diffcult is the jun/sep/dec/feb test? any patterns?

Feb tends to have a more forgiving curve for whatever reason.

But honestly- a more forgiving curve just means more challenging questions. It really does all even out in the end. Apply based on what's best for the admissions cycle and your prep process. Don't try to target a particular time of the year just because you think it may be easier/harder. The LSAT's pre-equating ALWAYS compensates for that.

I agree with the bottom part, but there is zero data to support your claim about feb scales. The most recent known feb scale is from the 2000 test, before LSAC started using the automated LSAT assembler software to produce test forms, so we have zero data about the scales of the last 14 Feb tests or any of them produced by the assembler program to current test specs. They could be super tight curves and easier tests for all we know, but more likely a fairly random mix where everything balances out anyway, making the whole issue really a moot point.

Take the test when you're ready to perform well and only then and don't worry about things outside your control, LSAC makes sure everything balances out so scaled scores properly represent the same skill/performance level no matter which test form it was achieved on.

You are correct that released data for Feb stopped at 2000. But there are benefits to being a massive prep company with thousands of student-reported data sets. From what I gathered while I worked for one (disclaimer- I was never someone who had direct access to that data- so I only heard it second-hand), this was true of pre-2000, and student-reported data showed that it continued to be true afterwards. But it was far from perfect data, so take it with a grain of salt.

Oh come'on KDL, having been an LSAT instructor you should know that there is no logically/mathematically possible way to determine or guesstimate the scale/curve of any single let alone a group of non disclosed test-forms from students achieved scores and answers they give in those highly subjective 'rate the test' surveys Kaplan tries to get students fill out after test day by comparing that data to their proctored diagnostic and practice test scores on other PTs taken as part of/while taking a prep class. Nice attempt at trying to get people to appeal to authority to believe your nonsensical/made up claim and also trying to distance yourself from it a bit now, lol.

Student reported test day scores are useful for prep companies to track student improvement rates and other stuff, but achieved scores data sets give zero information that could be used to determine or even guesstimate with any decent chance of reliability the scoring scale of any test-form you only have achieved scaled scores and some subjective survey results data about. At best student reports can sometimes indicate an unusually hard LG section which naturally would loosen up the scale a little and produce accurate predictions about the scale. However, there have been far more 'false alarms'/false positives where most people thought LGs or something about the test was extra hard and that the scale would be generous only to be disappointed when scores were released with disclosed tests and the scale turned out to be normal or tight compared to instances when the consensus that the scale would be more generous than usual turned out to be correct. I used to work for a big prep company for many years too and am familiar with what can and cannot be at least somewhat reliably figured out from students PT scores data sets, reported test day scores, students opinions of test difficulty right after taking it, and the purposes companies use the data for. Generalizing about trends in non-disclosed tests isn't one of them (at least not for people knowledgeable about the LSAT and skilled with logic) since it's clearly a pure guessing game with no way to verify accuracy of conclusions/guesstimates.

PS: All Feb tests up to 2000 except Feb 1998 are disclosed along with the scales. Of those eight tests, only four of them have generous scales (PT 7, Feb '97, PT B and PT C), the other four have normal to tight scales (-9 to -11 for 170), so what you said above about pre-2000 Feb tests/your companys data/conclusions about them is actually factually wrong.