What does 172+ really show? Forum

[Hey-O]

I'm not sure where to put this, but it is something I've been wondering about and I'm curious what other people think.

I know that law schools put a lot of emphasis on each extra point, but I don't think it really matters in terms of capabilities. I know that there are some consistent 175+ test takers but I really think that most people can consistently score above 170, but after that a lot of it is luck and test day performance. My range was very much 172-179 (never got a 180) and think this is pretty common for 170+ test takers.

IMHO I think between 172-180 it is really a wash (consistent 180ish outliers aside). It that most people in the 170s could score anywhere in the 170s. Any thoughts? Any data on how the extra points translate in law school success?

[JWicker10]

I agree with you that a lot of it in this range can be partially attributed to luck and raw test day performance, but those extra points make a huge difference in terms of how well your cycle will go. A 172 is not regarded the same as a 174, and a 174 isn't equal to a 175 and so on.

From what I've casually observed, anything over a 177 is pretty much considered a wash, but that's definitely not the case in the lower 170s. Applying to law school is very competitive and schools need a way to differentiate candidates. Moreover, average LSAT makes a difference for schools in USNWR rankings which is very significant.

Also, this is just my opinion, but lets say two people who are both scoring in the same range, 172-176. If one ends up with a 175 and the other a 171, it definitely shows at least to some extent how well the two individuals perform under pressure. While some is luck etc., there's also a performance factor that you should take into account. Maybe both got stuck on a few questions in one of the early sections. One person panicked a little and as a result missed a few extra questions. The other remained calm/relaxed/focused and the results speak for themselves.

[094320]

..

[Woozy]

I know that law schools put a lot of emphasis on each extra point, but I don't think it really matters in terms of capabilities.

It's well known that they do this to improve their own rankings.

IMHO I think between 172-180 it is really a wash (consistent 180ish outliers aside). It that most people in the 170s could score anywhere in the 170s. Any thoughts?

I think a person's LSAT proficiency is best approximated by a mean and a standard deviation. Consider the case of two testers, one with a mean/st. dev. of 175/2 and another with 172/3. In the course of many tests they are both likely to cover most of the 170 range, as you argue. However, this does not in any way mean that there is not a significant LSAT proficiency gap between them.

Now you will probably object that although there is a proficiency gap, the odds that the less proficient tester will outscore the more proficient one on a given administration are significant, thus admissions people should not place undue weight on the score difference. This is true for a singular case, however, admission committees are looking at a pool of applicants. While we cannot be too confident that any given 175 scorer is more LSAT- proficient than any given 172 scorer, we can be extremely confident that a group of 50 175 scorers are more proficient on the whole than a group of 50 172 scorers.

Whether this LSAT proficiency translates into law school success is a big debate. LSAC data supports the claim that it is the best single predictor of 1L grades, but not as good a predictor as LSAT/UG GPA combined.

[Audio Technica Guy]

yes and no. Yes, many people vary pretty wildly. However, most of that variance comes from poor pacing and strategy. For most people, if they stick to pacing and strategy (granted, that is easier said than done) you won't vary much more than 4 points. Most people's accuracy doesn't vary much more than 7% on a given test, if they don't have to rush due to poor pacing.

[09042014]

I wonder what the predictive ability of the LSAT is when you get to high scores. I wouldn't surprise me if above a certain LSAT, it drops off significantly.

[Audio Technica Guy]

I wonder what the predictive ability of the LSAT is when you get to high scores. I wouldn't surprise me if above a certain LSAT, it drops off significantly.

The tails of any normal distributed predictor aren't very good because they're so small. That is, there just aren't enough super high test takers to even gauge their predictability.

[KibblesAndVick]

I've always thought that if you score above 170 you should have to take a 2nd test. This test would be in the same format as the LSAT but much much harder. It wouldn't have any of the "easy" questions that most high scorers get correct. Every question would be hard as shit. The idea being that instead of having all of the really high scorers balled up within a 10 point rage you could get them to spread out into a nice normal distribution.

Also, you should be encouraged to take the LSAT multiple times instead of discouraged from doing so. That would create more data points for each test taker and center their sample LSAT score band closer to their "true" underlying score.

[094320]

..

[JD-INTIALSANDNAME]

I've always thought that if you score above 170 you should have to take a 2nd test. This test would be in the same format as the LSAT but much much harder. It wouldn't have any of the "easy" questions that most high scorers get correct. Every question would be hard as shit. The idea being that instead of having all of the really high scorers balled up within a 10 point rage you could get them to spread out into a nice normal distribution.

Also, you should be encouraged to take the LSAT multiple times instead of discouraged from doing so. That would create more data points for each test taker and center their sample LSAT score band closer to their "true" underlying score.

Interesting point, and I agree that people should take it more than once. Multiple data points are always more telling.

As interesting as I think it would be, I don't think there would be much practical purpose served by making people take another test. It seems as if very few people with well composed personal statements, strong academic records, and LSAT scores inside the 99% ever get denied from a top law school, including T5 schools.

[Audio Technica Guy]

I've always thought that if you score above 170 you should have to take a 2nd test. This test would be in the same format as the LSAT but much much harder. It wouldn't have any of the "easy" questions that most high scorers get correct. Every question would be hard as shit. The idea being that instead of having all of the really high scorers balled up within a 10 point rage you could get them to spread out into a nice normal distribution.

Also, you should be encouraged to take the LSAT multiple times instead of discouraged from doing so. That would create more data points for each test taker and center their sample LSAT score band closer to their "true" underlying score.

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.

[KibblesAndVick]

I've always thought that if you score above 170 you should have to take a 2nd test. This test would be in the same format as the LSAT but much much harder. It wouldn't have any of the "easy" questions that most high scorers get correct. Every question would be hard as shit. The idea being that instead of having all of the really high scorers balled up within a 10 point rage you could get them to spread out into a nice normal distribution.

Also, you should be encouraged to take the LSAT multiple times instead of discouraged from doing so. That would create more data points for each test taker and center their sample LSAT score band closer to their "true" underlying score.

Wait...why?
There aren't that many of us to begin with...

Well think of it this way. You have three test takes. Kid A is a super genius. He got every single question on the LSAT correct and finished every section with 5-10 minutes to spare. He got a 180. Kid B is also very smart and studied extensively. He missed 3 or 4 questions on the test but guessed on 5 or 6. He got a 178. Kid C is also very smart but a slight cut below the other two. He missed 8 questions but guessed on 10 or so. He got a 175. All of those scores are so close together that it's hard to differentiate. LSAT has no idea how many questions you guessed on or if you would be capable of doing just as well if the test was harder.

Now, lets say they all have to take the HardAsFuck LSAT (my personal choice of names). Kid A is still a super genius and get's every question right. 180. Kid B now misses, lets say, 15-20 questions. He gets a 160. Kid C misses 30-40 questions and gets a 150 or something. It allows the small differences between LSAT test takers to become exaggerated. As such, you can tell who actually has the most ability. It takes away the advantages of good guessing on test day, etc.

[KibblesAndVick]

I've always thought that if you score above 170 you should have to take a 2nd test. This test would be in the same format as the LSAT but much much harder. It wouldn't have any of the "easy" questions that most high scorers get correct. Every question would be hard as shit. The idea being that instead of having all of the really high scorers balled up within a 10 point rage you could get them to spread out into a nice normal distribution.

Also, you should be encouraged to take the LSAT multiple times instead of discouraged from doing so. That would create more data points for each test taker and center their sample LSAT score band closer to their "true" underlying score.

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.

I concede that this is probably the part of it that appeals most to me. I love data analysis and standardized testing. I'd rather work for LSAT than be a lawyer but what the hey.

[Woozy]

Well think of it this way. You have three test takes. Kid A is a super genius. He got every single question on the LSAT correct and finished every section with 5-10 minutes to spare. He got a 180. Kid B is also very smart and studied extensively. He missed 3 or 4 questions on the test but guessed on 5 or 6. He got a 178. Kid C is also very smart but a slight cut below the other two. He missed 8 questions but guessed on 10 or so. He got a 175. All of those scores are so close together that it's hard to differentiate. LSAT has no idea how many questions you guessed on or if you would be capable of doing just as well if the test was harder.

Now, lets say they all have to take the HardAsFuck LSAT (my personal choice of names). Kid A is still a super genius and get's every question right. 180. Kid B now misses, lets say, 15-20 questions. He gets a 160. Kid C misses 30-40 questions and gets a 150 or something. It allows the small differences between LSAT test takers to become exaggerated. As such, you can tell who actually has the most ability. It takes away the advantages of good guessing on test day, etc.

I think everyone knew what you were proposing. It's just that none of the consumers of the LSAT, students or schools, need this information.

[094320]

..

[KibblesAndVick]

I think everyone knew what you were proposing. It's just that none of the consumers of the LSAT, students or schools, need this information.

I see why most students wouldn't be interested but I don't follow the logic of why the schools wouldn't be interested. It seems like something of this sorts would allow them to better gauge whatever it is about the LSAT (predictive ability I guess) that makes them favor higher scores. I understand that a huge part of why they favor high scores is simply for rankings but there has to be something else there, no?

[Audio Technica Guy]

Well think of it this way. You have three test takes. Kid A is a super genius. He got every single question on the LSAT correct and finished every section with 5-10 minutes to spare. He got a 180. Kid B is also very smart and studied extensively. He missed 3 or 4 questions on the test but guessed on 5 or 6. He got a 178. Kid C is also very smart but a slight cut below the other two. He missed 8 questions but guessed on 10 or so. He got a 175. All of those scores are so close together that it's hard to differentiate. LSAT has no idea how many questions you guessed on or if you would be capable of doing just as well if the test was harder.

Now, lets say they all have to take the HardAsFuck LSAT (my personal choice of names). Kid A is still a super genius and get's every question right. 180. Kid B now misses, lets say, 15-20 questions. He gets a 160. Kid C misses 30-40 questions and gets a 150 or something. It allows the small differences between LSAT test takers to become exaggerated. As such, you can tell who actually has the most ability. It takes away the advantages of good guessing on test day, etc.

Being this smart on the LSAT doesn't help being a lawyer or law student at all. My best friend in law school got a 172 and was #1 in our class, is now clerking for SCOTUS and had guys from Wachtell and Williams and Connolly fighting over him.

[Audio Technica Guy]

I think everyone knew what you were proposing. It's just that none of the consumers of the LSAT, students or schools, need this information.

I see why most students wouldn't be interested but I don't follow the logic of why the schools wouldn't be interested. It seems like something of this sorts would allow them to better gauge whatever it is about the LSAT (predictive ability I guess) that makes them favor higher scores. I understand that a huge part of why they favor high scores is simply for rankings but there has to be something else there, no?

I don't think after you get past like 174 it matters very much at all, unless you prepped extensively to get that 174. I mean I can get a 180 using only 25 mins per section most of the time (in fact I did that on the last released LSAT, because I did it while proctoring an SAT). I wasn't particularly awesome in law school and during my brief time in a firm, I wasn't particularly noteworthy there either.

[romothesavior]

I wonder what the predictive ability of the LSAT is when you get to high scores. I wouldn't surprise me if above a certain LSAT, it drops off significantly.

I agree. The difference between someone with a 150 vs. someone with a 160 is a lot bigger than 170 vs. 180.

[KibblesAndVick]

People can miss 8 and get a 175?! Differentiating us seems like a waste of time and money considering how few of us there are. There aren't enough of us for schools to care or enough schools that are selective enough to afford to care. Esp. since law schools care about US News Ranking, which also doesn't care. Columbia doesn't care if a candidate is Kid A or Kid B, if the two have good GPAs, Columbia will take both and probably give both $.

I guess you're right that 8/175 isn't correct. But you get the gist. But what if the two candidates to Columbia don't have very high GPAs. Lets say that one majored in physics and got a 3.3 cumulative GPA. The other got a 3.6 as a business major. Both scored 175 on the LSAT. If you were Columbia wouldn't you be interested in how they did on a harder test that exaggerated their differences? If the physics major can still rock it on a really hard test and the business major can't doesn't that say something?

[KibblesAndVick]

I think everyone knew what you were proposing. It's just that none of the consumers of the LSAT, students or schools, need this information.

I see why most students wouldn't be interested but I don't follow the logic of why the schools wouldn't be interested. It seems like something of this sorts would allow them to better gauge whatever it is about the LSAT (predictive ability I guess) that makes them favor higher scores. I understand that a huge part of why they favor high scores is simply for rankings but there has to be something else there, no?

I don't think after you get past like 174 it matters very much at all, unless you prepped extensively to get that 174. I mean I can get a 180 using only 25 mins per section most of the time (in fact I did that on the last released LSAT, because I did it while proctoring an SAT). I wasn't particularly awesome in law school and during my brief time in a firm, I wasn't particularly noteworthy there either.

I'm assuming from your post that you took the LSAT, went to law school/firm, and then became an LSAT teacher and learned how to consistently score ~180 in that order? Did you score extremely high on the LSAT when you took it the first time?

[094320]

..

[Hey-O]

I wonder what the predictive ability of the LSAT is when you get to high scores. I wouldn't surprise me if above a certain LSAT, it drops off significantly.

I agree. The difference between someone with a 150 vs. someone with a 160 is a lot bigger than 170 vs. 180.

That is really the point I was trying to get at. I would be interested in the top 1% how predictive the LSAT really is as a measure of success in law school/being a lawyer. Schools seem to put a lot of weight on whether a person scores a 173 or a 176, but is there really that big of difference in their abilities?

I wonder if this isn't the reason that the HYS floor is around 170, because scoring above that range is what really matters and not how high you score above it.

[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.

[KibblesAndVick]

I guess you're right that 8/175 isn't correct. But you get the gist. But what if the two candidates to Columbia don't have very high GPAs. Lets say that one majored in physics and got a 3.3 cumulative GPA. The other got a 3.6 as a business major. Both scored 175 on the LSAT. If you were Columbia wouldn't you be interested in how they did on a harder test that exaggerated their differences? If the physics major can still rock it on a really hard test and the business major can't doesn't that say something?

No. Not necessarily. Depends on the validity of the test. Creating valid tests is difficult.

edited to not fuck up the quote, but i'll edit it again because i still fucked it up....

I agree that it's difficult. But do you mean to suggest that LSAC would be unable to do it?

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.