180s and percentile ranks Forum

[febstriver]

is it not true (based on the way percentile ranks work) that if one were to answer all questions correct on the LSAT then one would be in the 100th percentile?

whats even more puzzling (if my assumptions so far are true) is that here* Robin Singh claims to have scored all questions correct on the october 2003 administration but his percentile rank displays 99th. shouldn't be 100th if he got all questions correct?


http://www.testmasters.net/lsat/whyWeAr ... ction.aspx

[DoubleChecks]

180 is 99.9%

have you never taken standardized tests before? highest is always 99% or 99.9% lol

[Cosmo Kramer]

if you're in the 99th percentile, it means you scored better than 99 percent of test takers. if you're in the 100th percentile, that means you scored higher than 100 percent of test takers. That means that not only did NO ONE else get all questions correct, but you also that you scored higher than...yourself

[chocho]

if you score a 180 you dont outscore 100% of the people. other people score that as well. You outscore 99.9 though.

[Vincent Vega]

if you score a 180 you dont outscore 100% of the people. other people score that as well. You outscore 99.9 though.

titcr

[Nom Sawyer]

if you score a 180 you dont outscore 100% of the people. other people score that as well. You outscore 99.9 though.

titcr

even if u outscored every other person... u don't outscore yourself...

hence its an impossibility to be 100% rank

[jonnodotsg]

.

[febstriver]

what if i am the only person who got 100 or 101 questions correct on any particular administration? what then? still 99th?

[febstriver]

if you score a 180 you dont outscore 100% of the people. other people score that as well. You outscore 99.9 though.

titcr

even if u outscored every other person... u don't outscore yourself...

hence its an impossibility to be 100% rank

nvm

see above

[jonnodotsg]

.

[Zapatero]

what if i am the only person who got 100 or 101 questions correct on any particular administration? what then? still 99th?

Don't worry, you won't.

[HiLine]

Even though I agree that you cannot outscore yourself, hence achieve a 100th percentile rank, I simply find it puzzling that standardized test makers do not use mathematical logic in rounding up numbers.

[watts]

Even though I agree that you cannot outscore yourself, hence achieve a 100th percentile rank, I simply find it puzzling that standardized test makers do not use mathematical logic in rounding up numbers.

hehe

[eudaimondaimon]

Even though I agree that you cannot outscore yourself, hence achieve a 100th percentile rank, I simply find it puzzling that standardized test makers do not use mathematical logic in rounding up numbers.

Standardized tests are statistical creatures. Statisticians, who deal with confidence intervals all the time, feel right at home with numbers like .99995 and .999999.

[kittenmittons]

what if i am the only person who got 100 or 101 questions correct on any particular administration? what then? still 99th?

Don't worry, you won't.

[Flanker1067]

Even though I agree that you cannot outscore yourself, hence achieve a 100th percentile rank, I simply find it puzzling that standardized test makers do not use mathematical logic in rounding up numbers.

Standardized tests are statistical creatures. Statisticians, who deal with confidence intervals all the time, feel right at home with numbers like .99995 and .999999.

These types of numbers don't even apply to the LSAT. A 180 is 99.98%, resulting in anywhere from 15 to 25 180's a year depending on the amount of takers.

Edit: Of course, sometimes, you can apply these numbers, if you break it down even further. I am just saying, the LSAT is not built to determine who is in the 99.99999 percentile For that, the numbers of problems would have to be much greater, in order to seperate even those who are very good. (or bad)

[Shrimps]

Are you sure there are so few of them? I thought about 100-200 people score 180 annually.

[eudaimondaimon]

Even though I agree that you cannot outscore yourself, hence achieve a 100th percentile rank, I simply find it puzzling that standardized test makers do not use mathematical logic in rounding up numbers.

Standardized tests are statistical creatures. Statisticians, who deal with confidence intervals all the time, feel right at home with numbers like .99995 and .999999.

These types of numbers don't even apply to the LSAT. A 180 is 99.98%, resulting in anywhere from 15 to 25 180's a year depending on the amount of takers.

Edit: Of course, sometimes, you can apply these numbers, if you break it down even further. I am just saying, the LSAT is not built to determine who is in the 99.99999 percentile For that, the numbers of problems would have to be much greater, in order to seperate even those who are very good. (or bad)

Acknowledged. My intention wasn't to refer to the LSAT's percentiles specifically. I was just responding to HiLine's comment; my point being that numbers of that nature are not unusual when dealing with statistics and for the sake of precision they are almost NEVER rounded.

[Flanker1067]

Are you sure there are so few of them? I thought about 100-200 people score 180 annually.

I am not 100% sure as I don't work for/talk to the LSAC. This is just based on info from people who have looked into it, probably gotten 180's themselves, and I did neither of those. I don't get the impression that 100 people a year get a 180 at all though. Think about that, Yale could theoretically make a class entirely of 180's. I don't believe this to be true.

[Flanker1067]

Even though I agree that you cannot outscore yourself, hence achieve a 100th percentile rank, I simply find it puzzling that standardized test makers do not use mathematical logic in rounding up numbers.

Standardized tests are statistical creatures. Statisticians, who deal with confidence intervals all the time, feel right at home with numbers like .99995 and .999999.

These types of numbers don't even apply to the LSAT. A 180 is 99.98%, resulting in anywhere from 15 to 25 180's a year depending on the amount of takers.

Edit: Of course, sometimes, you can apply these numbers, if you break it down even further. I am just saying, the LSAT is not built to determine who is in the 99.99999 percentile For that, the numbers of problems would have to be much greater, in order to seperate even those who are very good. (or bad)

Acknowledged. My intention wasn't to refer to the LSAT's percentiles specifically. I was just responding to HiLine's comment; my point being that numbers of that nature are not unusual when dealing with statistics and for the sake of precision they are almost NEVER rounded.

Yea you are right, I am not judging or criticizing what you said. Just noting that we don't need a real statistician to handle the percentiles and stats regarding the LSAT.

[HiLine]

Even though I agree that you cannot outscore yourself, hence achieve a 100th percentile rank, I simply find it puzzling that standardized test makers do not use mathematical logic in rounding up numbers.

Standardized tests are statistical creatures. Statisticians, who deal with confidence intervals all the time, feel right at home with numbers like .99995 and .999999.

What if the precise percentile rank were 99.99999% and the test makers have to round the percentage up to 2 digits after decimal point for matter of convenience?

[tomwatts]

I don't get the impression that 100 people a year get a 180 at all though. Think about that, Yale could theoretically make a class entirely of 180's. I don't believe this to be true.

I was under the impression that Yale could make a class mostly (though perhaps not entirely) of 180's if it felt like it, but there are too many people with good GPAs and good softs who are a little under that that they don't. Yale and the like do deny people with 180's sometimes.

Oh, and some test-making companies do round naturally. When I got a perfect score on DAT Math, I was told that that was the 100.0th percentile, which is clearly impossible, but they were clearly rounding 99.96th percentile or above (because their precision was to the tenth's place). I don't believe that LSAC would be able to live with itself if it gave out a 100.0th percentile, though, because that's just so statistically wrong, and further, I think that a 180 is a 99.94th or lower percentile anyway.

[Flanker1067]

I don't get the impression that 100 people a year get a 180 at all though. Think about that, Yale could theoretically make a class entirely of 180's. I don't believe this to be true.

I was under the impression that Yale could make a class mostly (though perhaps not entirely) of 180's if it felt like it, but there are too many people with good GPAs and good softs who are a little under that that they don't. Yale and the like do deny people with 180's sometimes.

Sometimes they do, but if there were a 100 180's, then they would only have to let in 25% of them to have a 75th percentile of 180. They aren't even close to that, I think it is 177, and I highly doubt they reject 75% of 180's.

[Flanker1067]

Oh, and not that this is necessarily accurate, but LSN for last year says that of 9 180 applicants to Yale, 4 were accepted, 4 rejected, and a waitlist who didn't update.

[dutchstriker]

I'm pretty sure 180 is the 99.98 percentile. That's what it was in years past. If that's correct, some math shows it's unlikely that there are any more than 50 per year. Probably somewhere around 30-40.