What does 172+ really show? Forum

[094320]

..

[09042014]

I'm also of the opinion that after somewhere around 172 or 173 differences in LSAT scores don't really mean much. But I don't think that's because all of those people are of equal intelligence/ability. I think there are differences there but the LSAT has a ceiling effect after the low 170's.

The question is do law school exams have a similar ceiling effect.

[094320]

..

[BaiAilian2013]

This is the idea behind the confidence interval, right?

[KibblesAndVick]

I'm also of the opinion that after somewhere around 172 or 173 differences in LSAT scores don't really mean much. But I don't think that's because all of those people are of equal intelligence/ability. I think there are differences there but the LSAT has a ceiling effect after the low 170's.

The question is do law school exams have a similar ceiling effect.

Well unlike the LSAT I've never taken one. So feel free to disregard everything I'm about to write as hearsay.

My impression is that they would not because they are typically written exams instead of multiple choice (even though some use multiple choice that's not really the bread and butter). On a multiple choice test you can guess. It also imposes a mathematical limit on your score. You can't score better than 180. If you're writing an essay you can spot every issue. But even two people who spot "every" issue would not have written the same thing. Obviously this becomes subjective and they both probably get A's. What I'm trying to get at is that there's not the same sort of ceiling on a written examination.

Again, I'm not actually in any position to speak but why let that stop me.

[Hey-O]

This is the idea behind the confidence interval, right?

Do you mean the score band?

[sundance95]

I'm not really sure what "problem" you're trying to solve here. The LSAT is for law schools, and I can guarantee you they don't want that. I can guarantee you that 170+ people don't want that. Maybe a few statisticians might want this, and LSAC probably wouldn't mind a little extra income, but that's about it.

+1

[09042014]

I'm also of the opinion that after somewhere around 172 or 173 differences in LSAT scores don't really mean much. But I don't think that's because all of those people are of equal intelligence/ability. I think there are differences there but the LSAT has a ceiling effect after the low 170's.

The question is do law school exams have a similar ceiling effect.

It's not like professors use statisticians when constructing law school exams to make sure the exams are valid and consistent. It was this realization early in my undergrad career that caused my disillusionment with the professor made the exam the night before process.

I'm not sure it's possible to create a consistent and valid test for such a small group of people. Also consistent isn't really a factor when you are grading on a curve.

[Audio Technica Guy]

This is the idea behind the confidence interval, right?

No, not exactly. The confidence interval, or score band, just says that if you retook the test, and performed to the same level, you'd 95% likely get in that band. It more or less just takes guessing into account.

[Hey-O]

This is the idea behind the confidence interval, right?

No, not exactly. The confidence interval, or score band, just says that if you retook the test, and performed to the same level, you'd 95% likely get in that band. It more or less just takes guessing into account.

Hmmm... I wonder if the score bands for the 170s are wider than bands for lower scorers.

[09042014]

This is the idea behind the confidence interval, right?

No, not exactly. The confidence interval, or score band, just says that if you retook the test, and performed to the same level, you'd 95% likely get in that band. It more or less just takes guessing into account.

Not guessing, but performance variance for all factors. I'd guess that guessing is a small factor. I never guessed on questions and my score still had sway.

[HiLine]

I gotta believe over-studying also destroys the validity of your LSAT score. Most people aren't taking 50 PTs and going over each with a fine tooth comb. They take 3-5 prep tests and rock that's that.

Yeah, but why stop there? Over-studying destroys the validity of any exam and test scores, by the same token.

[09042014]

I gotta believe over-studying also destroys the validity of your LSAT score. Most people aren't taking 50 PTs and going over each with a fine tooth comb. They take 3-5 prep tests and rock that's that.

Yeah, but why stop there? Over-studying destroys the validity of any exam and test scores, by the same token.

Most post wasn't an indictment of the LSAT, just an observation.

Though tests can be written to be less learn able.

[Hey-O]

I gotta believe over-studying also destroys the validity of your LSAT score. Most people aren't taking 50 PTs and going over each with a fine tooth comb. They take 3-5 prep tests and rock that's that.

Yeah, but why stop there? Over-studying destroys the validity of any exam and test scores, by the same token.

I don't get this. We're not talking about IQ test. We're talking about potential and sometimes it takes practice to reach one's potential. Would you say a top athlete was less capable just because he had to practice to reach his potential?

[KibblesAndVick]

This is the idea behind the confidence interval, right?

No, not exactly. The confidence interval, or score band, just says that if you retook the test, and performed to the same level, you'd 95% likely get in that band. It more or less just takes guessing into account.

Hmmm... I wonder if the score bands for the 170s are wider than bands for lower scorers.

I'd imagine it's the opposite because people who score highly are more likely to have studied extensively. They've already "maxed out" to an extent and that should make them more consistent. Low scorers probably have a lot of people that would benefit simply by taking the real LSAT. More people who don't take it seriously and come in hungover and what not. Also, if you score poorly you probably guessed on a bunch of questions. If you scored highly you probably only guessed on a few. That would tighten the band for an individual high scorer.

I'd really like to see that data though. Sadly I get excited by this kind of crap.

[romothesavior]

I gotta believe over-studying also destroys the validity of your LSAT score. Most people aren't taking 50 PTs and going over each with a fine tooth comb. They take 3-5 prep tests and rock that's that.

Yeah, but why stop there? Over-studying destroys the validity of any exam and test scores, by the same token.

One of the key differences is that people view the LSAT as an indicator of intelligence and capability to succeed in law school, whereas people view grades on exams (like law exams) as evidence of both intelligence AND hard work. It is hard for schools to really use the LSAT as an equalizer to gauge an applicant's law school potential when some people are taking it cold and others spent thousands of dollars on tutors, classes, prep materials, etc. and put in tons of hours.

[09042014]

I gotta believe over-studying also destroys the validity of your LSAT score. Most people aren't taking 50 PTs and going over each with a fine tooth comb. They take 3-5 prep tests and rock that's that.

Yeah, but why stop there? Over-studying destroys the validity of any exam and test scores, by the same token.

I don't get this. We're not talking about IQ test. We're talking about potential and sometimes it takes practice to reach one's potential. Would you say a top athlete was less capable just because he had to practice to reach his potential?

But the LSAT isn't he the race, it's the qualifier. Just because you are sitting on pole position doesn't mean some guy who started 15th isn't going to smoke your ass in the real race.

I wasn't speaking out against the use of LSAT in admissions, I was mostly talking about comparing yourself to your law school class. Remember the average person at your school probably didn't LSAT gun. Compare your 5th PT to their real thing if you want to ballpark how well you measure up.

[Audio Technica Guy]

This is the idea behind the confidence interval, right?

No, not exactly. The confidence interval, or score band, just says that if you retook the test, and performed to the same level, you'd 95% likely get in that band. It more or less just takes guessing into account.

Hmmm... I wonder if the score bands for the 170s are wider than bands for lower scorers.

I think they're the same exact number of points on every test for every score. It's based on the way they scale the test to begin with.

[romothesavior]

But the LSAT isn't he the race, it's the qualifier. Just because you are sitting on pole position doesn't mean some guy who started 15th isn't going to smoke your ass in the real race.

Blatant pro-Texas anti-Yale trolling.

[Hey-O]

I think he was talking in law schools, not among law schools.

[romothesavior]

I think he was talking in law schools, not among law schools.

I understand that. I was being facetious.

[09042014]

I think he was talking in law schools, not among law schools.

Ha, I thought his joke was about Texans enjoying motor sports.

[Hey-O]

I think he was talking in law schools, not among law schools.

I understand that. I was being facetious.

Oh, Internets, I suck at you.

[romothesavior]

I think he was talking in law schools, not among law schools.

Ha, I thought his joke was about Texans enjoying motor sports.

This too.

[HiLine]

One of the key differences is that people view the LSAT as an indicator of intelligence and capability to succeed in law school, whereas people view grades on exams (like law exams) as evidence of both intelligence AND hard work. It is hard for schools to really use the LSAT as an equalizer to gauge an applicant's law school potential when some people are taking it cold and others spent thousands of dollars on tutors, classes, prep materials, etc. and put in tons of hours.

Grades in history, physics, music, philosophy, etc. are also indicators of your ability to do your job well; so they are also indicators of intelligence and capability to succeed in their job.