Really dumb curve question

(Please Ask Questions and Answer Questions)
User avatar
zreinhar
Posts: 308
Joined: Wed May 27, 2009 12:15 pm

Really dumb curve question

Postby zreinhar » Thu Apr 14, 2011 12:14 pm

I am a stupid lowly 0L, who figured it would be quicker to ask.

The law school curve is similair to a bellcurve correct? In that around 50% of the students are at median? Or is it more linear in that, only one student is at median and a lot others are a percentage points away?

Edit for typos

User avatar
Law Sauce
Posts: 923
Joined: Mon Nov 15, 2010 2:21 pm

Re: Really dumb curve question

Postby Law Sauce » Thu Apr 14, 2011 12:23 pm

closer to a bell curve i believe, with a lot of students closer to median, though i would guess that there is a much longer tail on the upper end (ppl distinguishing themselves farther and farther out)

User avatar
zreinhar
Posts: 308
Joined: Wed May 27, 2009 12:15 pm

Re: Really dumb curve question

Postby zreinhar » Thu Apr 14, 2011 12:26 pm

Great, thanks.

Also, I don't know if you know this, or someone else might, when they rank for a PT program, and some people transfer full time is the ranking adjusted? i.e. if I was 3rd and the top two kids transfer, I would be 1st the following semester?

User avatar
Patriot1208
Posts: 7044
Joined: Tue May 18, 2010 11:28 am

Re: Really dumb curve question

Postby Patriot1208 » Thu Apr 14, 2011 12:29 pm

Wrong, it isn't a curve, it's a distribution. Meaning the median will be a certain grade, say a b, and then the whole class will be fit into that distribution. You are ranked 1 through whatever, and then the median person is put at whatever grade, and then a certain percentage will be assigned each letter grade.

840e
Posts: 47
Joined: Fri Jan 28, 2011 3:17 pm

Re: Really dumb curve question

Postby 840e » Thu Apr 14, 2011 12:32 pm

.
Last edited by 840e on Sat Nov 22, 2014 3:25 pm, edited 1 time in total.

User avatar
zreinhar
Posts: 308
Joined: Wed May 27, 2009 12:15 pm

Re: Really dumb curve question

Postby zreinhar » Thu Apr 14, 2011 12:34 pm

Patriot1208 wrote:Wrong, it isn't a curve, it's a distribution. Meaning the median will be a certain grade, say a b, and then the whole class will be fit into that distribution. You are ranked 1 through whatever, and then the median person is put at whatever grade, and then a certain percentage will be assigned each letter grade.


But doesnt that mean that many people will be at one position? Say at that B? And what about people who ahve the same grades or nearly identical?

User avatar
zreinhar
Posts: 308
Joined: Wed May 27, 2009 12:15 pm

Re: Really dumb curve question

Postby zreinhar » Thu Apr 14, 2011 12:35 pm

840e wrote:when i visited GW pt admitted students session, they said that pt students grades are considered within the entire 1L class in terms of rank. They don't distinguish that you might be number 1 or 80 within the part time program... it is how you fall in terms of the overall class. By the time you graduate, the people that transferred into the ft would be compared to the people graduating with them... and the people who remained in the pt program (and obviously took longer) would be compared with rank to the group/year of people they graduate with.



Oh wow, that sucks, so even though I have less time in the day/to dedicate I have to make the same grade the FT students do?

User avatar
Patriot1208
Posts: 7044
Joined: Tue May 18, 2010 11:28 am

Re: Really dumb curve question

Postby Patriot1208 » Thu Apr 14, 2011 12:37 pm

zreinhar wrote:
Patriot1208 wrote:Wrong, it isn't a curve, it's a distribution. Meaning the median will be a certain grade, say a b, and then the whole class will be fit into that distribution. You are ranked 1 through whatever, and then the median person is put at whatever grade, and then a certain percentage will be assigned each letter grade.


But doesnt that mean that many people will be at one position? Say at that B? And what about people who ahve the same grades or nearly identical?

sure but the distribution over multiple classes makes individuals distinguish themselves so it doesn't form like a curve would

840e
Posts: 47
Joined: Fri Jan 28, 2011 3:17 pm

Re: Really dumb curve question

Postby 840e » Thu Apr 14, 2011 12:41 pm

.
Last edited by 840e on Sat Nov 22, 2014 3:26 pm, edited 1 time in total.

User avatar
Always Credited
Posts: 2509
Joined: Wed Jun 24, 2009 1:31 pm

Re: Really dumb curve question

Postby Always Credited » Thu Apr 14, 2011 12:42 pm

zreinhar wrote:
840e wrote:when i visited GW pt admitted students session, they said that pt students grades are considered within the entire 1L class in terms of rank. They don't distinguish that you might be number 1 or 80 within the part time program... it is how you fall in terms of the overall class. By the time you graduate, the people that transferred into the ft would be compared to the people graduating with them... and the people who remained in the pt program (and obviously took longer) would be compared with rank to the group/year of people they graduate with.



Oh wow, that sucks, so even though I have less time in the day/to dedicate I have to make the same grade the FT students do?


At GW 9/10 of the PT students don't even work, so those who do are fucked. And those who don't dominate. Further, the PT students share two classes with FT students.

Just saying.

User avatar
zreinhar
Posts: 308
Joined: Wed May 27, 2009 12:15 pm

Re: Really dumb curve question

Postby zreinhar » Thu Apr 14, 2011 12:46 pm

Patriot1208 wrote:
zreinhar wrote:
Patriot1208 wrote:Wrong, it isn't a curve, it's a distribution. Meaning the median will be a certain grade, say a b, and then the whole class will be fit into that distribution. You are ranked 1 through whatever, and then the median person is put at whatever grade, and then a certain percentage will be assigned each letter grade.


But doesnt that mean that many people will be at one position? Say at that B? And what about people who ahve the same grades or nearly identical?

sure but the distribution over multiple classes makes individuals distinguish themselves so it doesn't form like a curve would


I was more referring to the all powerful 1L grades

User avatar
zreinhar
Posts: 308
Joined: Wed May 27, 2009 12:15 pm

Re: Really dumb curve question

Postby zreinhar » Thu Apr 14, 2011 12:48 pm

840e wrote:it was presented in terms of you have less courses to juggle in comparison to the ft students and more semesters to bring your grades up if need be... ( i guess or down... but i hope not)

they did stress that they put more emphasis on distinctions they put on your transcript (like top 30% get x award and top 10% get x award), but it is done for the class year as a whole...

it made me a little nervous because most examples of people who had good enough grades had extremely flexible jobs in terms of time/ part time jobs. As I have a full time job (that in an ideal world i would love to keep), it caused me a bit of concern of whether or not I would actually be able to remain competitive with my pt peers with more flexible work options. I haven't made my decision yet (need to by cob tomorrow), but it is definitely making me mull over my options.



Always Credited wrote:
zreinhar wrote:
840e wrote:when i visited GW pt admitted students session, they said that pt students grades are considered within the entire 1L class in terms of rank. They don't distinguish that you might be number 1 or 80 within the part time program... it is how you fall in terms of the overall class. By the time you graduate, the people that transferred into the ft would be compared to the people graduating with them... and the people who remained in the pt program (and obviously took longer) would be compared with rank to the group/year of people they graduate with.



Oh wow, that sucks, so even though I have less time in the day/to dedicate I have to make the same grade the FT students do?


At GW 9/10 of the PT students don't even work, so those who do are fucked. And those who don't dominate. Further, the PT students share two classes with FT students.

Just saying.


All of that scares the crap out of me. I'm in at GMU, WL at GW and WL at GULC. My job is really flexible for time, which will help. But is very big on my production, if that falters, I am unemployed. Ah well, we will wait and see. I have to deposit at GMU anyways tomorrow, so I'll ask how they do it.

User avatar
Patriot1208
Posts: 7044
Joined: Tue May 18, 2010 11:28 am

Re: Really dumb curve question

Postby Patriot1208 » Thu Apr 14, 2011 1:28 pm

zreinhar wrote:
Patriot1208 wrote:
zreinhar wrote:
Patriot1208 wrote:Wrong, it isn't a curve, it's a distribution. Meaning the median will be a certain grade, say a b, and then the whole class will be fit into that distribution. You are ranked 1 through whatever, and then the median person is put at whatever grade, and then a certain percentage will be assigned each letter grade.


But doesnt that mean that many people will be at one position? Say at that B? And what about people who ahve the same grades or nearly identical?

sure but the distribution over multiple classes makes individuals distinguish themselves so it doesn't form like a curve would


I was more referring to the all powerful 1L grades

Yes they are fit to a distribution, not curved

User avatar
Wholigan
Posts: 763
Joined: Sat Jan 29, 2011 4:51 pm

Re: Really dumb curve question

Postby Wholigan » Thu Apr 14, 2011 7:12 pm

.
Last edited by Wholigan on Fri Apr 15, 2011 9:43 pm, edited 1 time in total.

User avatar
drdolittle
Posts: 628
Joined: Tue Mar 09, 2010 4:15 am

Re: Really dumb curve question

Postby drdolittle » Thu Apr 14, 2011 7:40 pm

Patriot1208 wrote:Wrong, it isn't a curve, it's a distribution. Meaning the median will be a certain grade, say a b, and then the whole class will be fit into that distribution. You are ranked 1 through whatever, and then the median person is put at whatever grade, and then a certain percentage will be assigned each letter grade.

Not wrong. You've just described what a curve represents in layman's terms. It's a distribution graphed. A continuous probably (normal) distribution graphed forms a bell shaped curve. Distribution / curve are effectively synonymous in this context.

But your description of how letter grades are practically assigned is right on.

User avatar
Lawquacious
Posts: 2037
Joined: Fri Oct 23, 2009 10:36 am

Re: Really dumb curve question

Postby Lawquacious » Thu Apr 14, 2011 7:56 pm

drdolittle wrote:
Patriot1208 wrote:Wrong, it isn't a curve, it's a distribution. Meaning the median will be a certain grade, say a b, and then the whole class will be fit into that distribution. You are ranked 1 through whatever, and then the median person is put at whatever grade, and then a certain percentage will be assigned each letter grade.

Not wrong. You've just described what a curve represents in layman's terms. It's a distribution graphed. A continuous probably (normal) distribution graphed forms a bell shaped curve. Distribution / curve are effectively synonymous in this context.


+1

Also, I don't think that all law schools necessary determine their 'curve' in the same way. But it is true that the curve/distribution of law school grades is generally different from a 'curve' at the UG level; in UG I believe the distribution or allocation of grades can be set to one individual's performance (usually a high scorer for the class can 'set' the curve if there is one). At least at my law school, the curve/distribution is pre-set, so that a certain percentage of students fall into a certain range, regardless of what the individual data points are. Also, the distribution isn't determined by the median as Patriot indicated, but rather the median is determined by the distribution (which is pre-set, except there is a 2.5 percent margin of error for the number of grades that fall in a paricular range allowed, and this in turn means that the median isn't going to stay constant at my school). I could be wrong about any or all of this (and am willing to stand corrected by someone who can cogently explain otherwise), but at least as far as my school goes I think this is accurate. It may be largely or primarily a matter of different usage though (semantics).

User avatar
Patriot1208
Posts: 7044
Joined: Tue May 18, 2010 11:28 am

Re: Really dumb curve question

Postby Patriot1208 » Thu Apr 14, 2011 9:44 pm

drdolittle wrote:
Patriot1208 wrote:Wrong, it isn't a curve, it's a distribution. Meaning the median will be a certain grade, say a b, and then the whole class will be fit into that distribution. You are ranked 1 through whatever, and then the median person is put at whatever grade, and then a certain percentage will be assigned each letter grade.

Not wrong. You've just described what a curve represents in layman's terms. It's a distribution graphed. A continuous probably (normal) distribution graphed forms a bell shaped curve. Distribution / curve are effectively synonymous in this context.

But your description of how letter grades are practically assigned is right on.

I see we are coming up with our own definitions of terms now. But no, you are still wrong. When something is fit to a distribution that has a determined percentage of students with certain grades it doesn't form a normal curve. Depending on the distribution it will look much more linear than a bell curve would. Therefore, that is still wrong.

ETA Think of it this way. When a curve is implemented everyones grades are just changed a numerical amount. This just keeps the normal distribution but moves it to a different median. In a distribution the persons are ranked linearly and then assigned ranks 1-whatever which turns into grade distributions. This doesn't adhere to a normal distribution that would look like a bell curve and have a normal standard deviation. Instead it is arbitrarily fitted in a way that does not reflect a normal distribution. Whether or not the laymen understand this it is still an important difference, especially since the laymen seems to think these curves are set on a bell curve with a normal distribution, which is not the case.

User avatar
drdolittle
Posts: 628
Joined: Tue Mar 09, 2010 4:15 am

Re: Really dumb curve question

Postby drdolittle » Thu Apr 14, 2011 10:55 pm

Patriot1208 wrote:
drdolittle wrote:
Patriot1208 wrote:Wrong, it isn't a curve, it's a distribution. Meaning the median will be a certain grade, say a b, and then the whole class will be fit into that distribution. You are ranked 1 through whatever, and then the median person is put at whatever grade, and then a certain percentage will be assigned each letter grade.

Not wrong. You've just described what a curve represents in layman's terms. It's a distribution graphed. A continuous probably (normal) distribution graphed forms a bell shaped curve. Distribution / curve are effectively synonymous in this context.

But your description of how letter grades are practically assigned is right on.

I see we are coming up with our own definitions of terms now. But no, you are still wrong. When something is fit to a distribution that has a determined percentage of students with certain grades it doesn't form a normal curve. Depending on the distribution it will look much more linear than a bell curve would. Therefore, that is still wrong.

ETA Think of it this way. When a curve is implemented everyones grades are just changed a numerical amount. This just keeps the normal distribution but moves it to a different median. In a distribution the persons are ranked linearly and then assigned ranks 1-whatever which turns into grade distributions. This doesn't adhere to a normal distribution that would look like a bell curve and have a normal standard deviation. Instead it is arbitrarily fitted in a way that does not reflect a normal distribution. Whether or not the laymen understand this it is still an important difference, especially since the laymen seems to think these curves are set on a bell curve with a normal distribution, which is not the case.

Not exactly sure what own definitions of terms you're referring to. And why am I still wrong? I was never wrong to begin with. 8) I was using the normal distribution as a quick & simple example of how a distribution can be represented graphically by a curve, a concept you seemed to be refuting by distinguishing between distribution and curve. But I get your point: law school final grades are typically not distributed normally on a bell curve. They're still the result of curving though.

The initial distribution of raw points earned in a class per student will likely fall on a normal distribution, or approximately that if the sample size is large enough. Of course you're right that the final distribution of grades does not necessarily have to conform to a normal distribution however. As you described, the final grading scheme will come up with a distribution of grades based on school/prof policies through "curving." Curving here is just a process of assigning grades based on some arbitrary percentile intervals (e.g., top 5% = A, next 10% = A-, etc...). And this percentile distribution of grades could once again be binned, plotted and then fitted by a probability function, just like the original distribution of raw points could be fit by a Gaussian, forming a curve, which would graphically reflect the final grade distribution, whatever it may be.

User avatar
Patriot1208
Posts: 7044
Joined: Tue May 18, 2010 11:28 am

Re: Really dumb curve question

Postby Patriot1208 » Fri Apr 15, 2011 9:47 am

drdolittle wrote:
Patriot1208 wrote:
drdolittle wrote:
Patriot1208 wrote:Wrong, it isn't a curve, it's a distribution. Meaning the median will be a certain grade, say a b, and then the whole class will be fit into that distribution. You are ranked 1 through whatever, and then the median person is put at whatever grade, and then a certain percentage will be assigned each letter grade.

Not wrong. You've just described what a curve represents in layman's terms. It's a distribution graphed. A continuous probably (normal) distribution graphed forms a bell shaped curve. Distribution / curve are effectively synonymous in this context.

But your description of how letter grades are practically assigned is right on.

I see we are coming up with our own definitions of terms now. But no, you are still wrong. When something is fit to a distribution that has a determined percentage of students with certain grades it doesn't form a normal curve. Depending on the distribution it will look much more linear than a bell curve would. Therefore, that is still wrong.

ETA Think of it this way. When a curve is implemented everyones grades are just changed a numerical amount. This just keeps the normal distribution but moves it to a different median. In a distribution the persons are ranked linearly and then assigned ranks 1-whatever which turns into grade distributions. This doesn't adhere to a normal distribution that would look like a bell curve and have a normal standard deviation. Instead it is arbitrarily fitted in a way that does not reflect a normal distribution. Whether or not the laymen understand this it is still an important difference, especially since the laymen seems to think these curves are set on a bell curve with a normal distribution, which is not the case.

Not exactly sure what own definitions of terms you're referring to. And why am I still wrong? I was never wrong to begin with. 8) I was using the normal distribution as a quick & simple example of how a distribution can be represented graphically by a curve, a concept you seemed to be refuting by distinguishing between distribution and curve. But I get your point: law school final grades are typically not distributed normally on a bell curve. They're still the result of curving though.

The initial distribution of raw points earned in a class per student will likely fall on a normal distribution, or approximately that if the sample size is large enough. Of course you're right that the final distribution of grades does not necessarily have to conform to a normal distribution however. As you described, the final grading scheme will come up with a distribution of grades based on school/prof policies through "curving." Curving here is just a process of assigning grades based on some arbitrary percentile intervals (e.g., top 5% = A, next 10% = A-, etc...). And this percentile distribution of grades could once again be binned, plotted and then fitted by a probability function, just like the original distribution of raw points could be fit by a Gaussian, forming a curve, which would graphically reflect the final grade distribution, whatever it may be.

I guess I miss read your first post. But the point being is that grades aren't assigned with a curve that fits under a normal distribution and a bell curve. That was the original question. Grades are instead assigned by a proportional distribution which will be much more linear than a bell curve.

User avatar
D-ROCCA
Posts: 324
Joined: Mon May 17, 2010 10:14 pm

Re: Really dumb curve question

Postby D-ROCCA » Fri Apr 15, 2011 10:05 am

Test scores are almost certainly not iid for individual classes. Even though there is a large sample size the raw scores almost certainly won't be normally distributed, they'll be skewed right for individual classes. The professor determines the distribution (the shape of the curve). When your grades for all your classes are aggregated, then those scores will most likely be normally distributed.

nigelfrost
Posts: 122
Joined: Tue Apr 20, 2010 8:42 pm

Re: Really dumb curve question

Postby nigelfrost » Fri Apr 15, 2011 10:12 am

Patriot1208 wrote:I guess I miss read your first post.


(1) Who's "Miss Read?"
(2) Where can I find her?

User avatar
D-ROCCA
Posts: 324
Joined: Mon May 17, 2010 10:14 pm

Re: Really dumb curve question

Postby D-ROCCA » Fri Apr 15, 2011 10:12 am

nigelfrost wrote:
Patriot1208 wrote:I guess I miss read your first post.


(1) Who's "Miss Read?"
(2) Where can I find her?


Clever, I see what you did there...

nigelfrost
Posts: 122
Joined: Tue Apr 20, 2010 8:42 pm

Re: Really dumb curve question

Postby nigelfrost » Fri Apr 15, 2011 10:13 am

D-ROCCA wrote:
nigelfrost wrote:
Patriot1208 wrote:I guess I miss read your first post.


(1) Who's "Miss Read?"
(2) Where can I find her?


Clever, I see what you did there...


I'm like the Penn & Teller of language.

User avatar
Patriot1208
Posts: 7044
Joined: Tue May 18, 2010 11:28 am

Re: Really dumb curve question

Postby Patriot1208 » Fri Apr 15, 2011 10:16 am

nigelfrost wrote:
Patriot1208 wrote:I guess I miss read your first post.


(1) Who's "Miss Read?"
(2) Where can I find her?

woops... early morning at work.




Return to “Law School FAQ”

Who is online

Users browsing this forum: No registered users and 2 guests