... Forum

[imperspective]

...

[Mal Reynolds]

I've been wondering, since the LSAT is learnable for (at least almost) everyone, do groups of people who provide moral and strategic support like TLS have a serious effect on the curve? How many people waking up to the fact that they can do better would it take for the curve to break down? Or, do I still fail to understand it despite all the lurking I've done?!

The LSAT isn't curved.

[aboutmydaylight]

There's no curve. The test is equated before its taken. That means there isn't a certain percentage of people that MUST get a certain score or above/below a certain score. Because psychometricians are good at their job though, it usually is this case that certain scores represent certain percentiles. In that regard, if you took out high competitive/high test score takers, it likely wouldn't affect the "curve" at all, at least not initially. In the long run though, LSAC might think their tests are getting harder because there is less high scorers, so that could affect their desicion making. Or they could keep it the same, and lower test scores would have more value since there is less higher test scores.

So yes, in that sense, TLS does affect it I guess.

[lawschool2014hopeful]

Most definitely.

People try to drag out the distinction between a curve and "equated" too much lolz.

LSAT is essentially a curved test based on past 3 years of data, rather than on a single test/examination.

So if people are getting better, scoring a certain score is harder.

For example, 170 use to be 99th percentile in early tests and now is the 97th. Also, test-takers can tell, scoring 170 in the most recent tests is definitely harder than an older test, both subjective feeling and objectively (less room for errors).

[rinkrat19]

Most definitely.

People try to drag out the distinction between a curve and "equated" too much lolz.

LSAT is essentially a curved test based on past 3 years of data, rather than on a single test/examination.

So if people are getting better, scoring a certain score is harder.

For example, 170 use to be 99th percentile in early tests and now is the 97th. Also, test-takers can tell, scoring 170 in the most recent tests is definitely harder than an older test, both subjective feeling and objectively (less room for errors).

(doesn't understand curves, percentiles, or equating)

[Pneumonia]

Most definitely.

People try to drag out the distinction between a curve and "equated" too much lolz.

LSAT is essentially a curved test based on past 3 years of data, rather than on a single test/examination.

So if people are getting better, scoring a certain score is harder.

For example, 170 use to be 99th percentile in early tests and now is the 97th. Also, test-takers can tell, scoring 170 in the most recent tests is definitely harder than an older test, both subjective feeling and objectively (less room for errors).

apart from 170 now being around the 97th, none of this is really true.

[lawschool2014hopeful]

I honestly dont see how I am wrong, please do explain.

This is a standardized test, if people are doing better, the test gets objectively harder.

Yes, test items are equated/scored on difficulty based on the experimental section.

If people are doing better on the experimental overall, i.e., more studying, the test has to be harder in order to achieve a higher score.

[Jeffort]

I honestly dont see how I am wrong, please do explain.

This is a standardized test, if people are doing better, the test gets objectively harder.

Standardized tests by definition are supposed to produce scores that represent an objective measurement of particular skill levels, not a bell curve distribution.

Each scaled score is meant to represent a particular objective skill/ability level so that scores from different test-forms/administrations are comparable even if the scores are achieved a few years apart. The percentile rank for each score slightly changes over the years due to minor variations in test taker populations, but the skill level each score represents and the difficulty in obtaining each score stays the same so that LSs can compare LSAT scores from different applicants that were achieved years apart from one another. It's the essence of being a standardized test. If LSAC set out to make getting 170+ harder, then a 170 would suddenly mean something different than it used to and not be comparable to a 170 from previous tests.

The complex statistical methods LSAC uses to equate test forms is designed to make sure that LSAT scores earned within a three year window of each other are comparable and represent the same actual skill/ability level.

[lawschool2014hopeful]

I honestly dont see how I am wrong, please do explain.

This is a standardized test, if people are doing better, the test gets objectively harder.

Each scaled score is meant to represent a particular objective skill/ability level so that scores from different test-forms/administrations are comparable even if the scores are achieved a few years apart. The percentile rank for each score slightly changes over the years due to minor variations in test taker populations, but the skill level each score represents and the difficulty in obtaining each score stays the same so that LSs can compare LSAT scores from different applicants that were achieved years apart from one another. It's the essence of being a standardized test.

Right I agree, while each score is comparable within the few year (i.e., 3 year comparison period), it has gotten harder to achieve a higher percentile score now than it was 6 years ago.

The idea of particular objective skill/ability is all stated in relative terms anyway. A question is hard not because it is logically fundamentally hard, it is because only higher scorers get it right consistently. Sure one could argue that is because it is reliant on a high level logical skill, but we would be going in a circle.

[Jeffort]

Oh, ok. Yeah, it's harder to be 99th % now since that is a higher scaled score than it used to be, I thought you meant they would make it harder to get each particular scaled score.

The circular nature of the test construction and equating process is interesting though, it kinda begs the old chicken vs the egg, which came first question.

[lawschool2014hopeful]

Oh, ok. Yeah, it's harder to be 99th % now since that is a higher scaled score than it used to be, I thought you meant they would make it harder to get each particular scaled score.

The circular nature of the test construction and equating process is interesting though, it kinda begs the old chicken vs the egg, which came first question.

I make both claims.

170 on Prep test 71 is most definitely more difficult than 170 on Preptest 10 in any sort of objective fashion you would want to measure (this is of course is more of a stretch claim, and perhaps totally unfounded), but a 170 on 71 is probably not much different from PT 50, even though a PT 50 a 170 would rep a higher percentile score.

Haha yea, the test equating process is quite interesting indeed.