Whats the curve in your class?

[djheater]

<--- read "What's the curve in your ass?"

{{looks around, gets bored, leaves thread}}

 

[beer]

Originally posted by: WhiteKnight
Here at Johns Hopkins most of the engineering courses are curved to a C.

Most of our grades at UT are curved to a C+. If you're 3% or more above average, you get a B.

 

[Mears]

Amish, I can't pretend to have read all the back and forth posts in this thread, but let me give you my take.

Jack goes to a university A

Jill goes to university B

University A and B are both highly regarded institutions

University A doesn't curve any exams even though they are much more difficult than the exams proctored at university B. Jack was the the graduated with a 2.5 which put him in the top 15% of his class.

Jill easily obtains a 4.0 with the lax academic atmosphere at university B.

Now when Jack goes out in the work force he is going to have to compete with all the Jills out there who have 4.0s on their resume even though he easily could have obtained a 4.0 at university B.

 

[TallBill]

Originally posted by: Elemental007

Originally posted by: Electric Amish

Why does it have to be certain percentages at any given section?

amish

By your logic and "reasoning," by the time students graduated, there wouldn't be any students left in the class. No one I have ever met in our EE program comes close to making a 70% in each and every class they make. Professors make different tests, some harder than others, and by forcing each professor to have so many As, so many Bs, etc etc, you can identify your good students from students that just had an easy professor.

And no, professors will not "standardize" tests. This isn't high school again.

Hehe, forget about students actually earning their accomplishments. Cuz it would look bad for the school if students didnt pass.

Btw mears, you are completely wrong

 

[XZeroII]

Not many. Usually, if there was one or two questions that were unusually difficult on a test and most people got it wrong, the points were just refunded and everyone gets an extra 2 points or something. Curves where 20% get A's 30% get B's etc., are stupid IMO. People should all be graded on the same scale against the same standards. I could take a class and get an 89% avg, yet get a C because there were over 50% of the other people got 90% or higher. If I had waited, the next semester I might get an A for getting the same grades. It's not consistent.

 

[OldSpooky]

Difficult exams are given for a reason - they are a stress test of a student's mastery of material. If an exam is easy and everyone got 90% +, there is no real way for the instructor to know who really knows their stuff and who doesn't. It would be like making a 3DMark test that allowed all GPUs to score 20,000.

I felt pretty awful getting a 50% on an algorithms exam a couple years ago, but I reminded myself that the questions I did answer I got right - I knew my stuff, I just didn't have the years of experience necessary to answer everything expediently. And the class average was about that, so no worries.

 

[dullard]

I think all of you are confusing two topics (a) scaling and (b) curving.

Scaling means that if a class/test was made so hard that no one could get into the A range, then everyone is raised so that some people get an A. (Or conversely if it was so easy that if everyone got in the A range, then the lower A's will be dropped down to a non-A grade.)

Curving means that a number (such as 10%) will fail no matter how well they do in the class. For example, if we made 100 of the world's top physics experts take a beginning level physics course, then with a curve 10 of them would be forced to fail no matter what. They could get every problem perfectly correct on every test and every homework assignment, but since there is a curve 10% must still fail. And in my opinion that is rediculous. Second example, if everyone does no work at all, no one does homework, no one gets anything correct on the tests, then still 10% get an A. Wow great for them aceing a class and not knowing anything at all!

These are two completely different things and would fix almost all arguments on this thread if you understood the difference.

 

[beer]

Originally posted by: dullard
I think all of you are confusing two topics (a) scaling and (b) curving.

Scaling means that if a class/test was made so hard that no one could get into the A range, then everyone is raised so that some people get an A. (Or conversely if it was so easy that if everyone got in the A range, then the lower A's will be dropped down to a non-A grade.)

Curving means that a number (such as 10%) will fail no matter how well they do in the class. For example, if we made 100 of the world's top physics experts take a beginning level physics course, then with a curve 10 of them would be forced to fail no matter what. They could get every problem perfectly correct on every test and every homework assignment, but since there is a curve 10% must still fail. And in my opinion that is rediculous. Second example, if everyone does no work at all, no one does homework, no one gets anything correct on the tests, then still 10% get an A. Wow great for them aceing a class and not knowing anything at all!

These are two completely different things and would fix almost all arguments on this thread if you understood the difference.

You are basing your opinion on curving on two extremes that never happen.

In my physics class (of 140 students), 1 student had a 95, 1 had a 90. 9 had an 80-90, and the rest were below the traditional 'B' range. By your defnition of 'scaling' you would just raise everyone's grade less than 5 points. In reality, it's a mixture of scaling and curving - but the two extremes you mentioned don't really count because they're not valid points.

 

[fizmeister]

I'm taking a math class next semester where only two As have been given in the past 10 years.

So...yeah.

 

[TallBill]

Originally posted by: OldSpooky
Difficult exams are given for a reason - they are a stress test of a student's mastery of material. If an exam is easy and everyone got 90% +, there is no real way for the instructor to know who really knows their stuff and who doesn't. It would be like making a 3DMark test that allowed all GPUs to score 20,000.

The 3dmark reference is way off. Whats the point of testing all the GPUS really hard, then boosting their score? or altering their direct comparison.


Also, i agree with Elemental007 that its a combination of scaling/curving/fitting grades in which is the issue.

 

[TallBill]

Originally posted by: fizmeister
I'm taking a math class next semester where only two As have been given in the past 10 years.

So...yeah.

And I bet that those 2 students were really smart. At least you know what your going into. No false expectations.

 

[dullard]

Originally posted by: Elemental007

You are basing your opinion on curving on two extremes that never happen.

In my physics class (of 140 students), 1 student had a 95, 1 had a 90. 9 had an 80-90, and the rest were below the traditional 'B' range. By your defnition of 'scaling' you would just raise everyone's grade less than 5 points. In reality, it's a mixture of scaling and curving - but the two extremes you mentioned don't really count because they're not valid points.

Teach a class for a few years and you will see that things like I mentioned happen all the time (but not quite to that extreme). For example, at my university there was a major push in 1994 to get honors students (in the form of a drastic increase in performance based scholarships and free books to the top students) to come to the university at the same time there was a strict increase in requirements (meaning thousands of borderline students were not accepted). Thus in 1995, a wave of students came that were on average far, far more qualified than in any other year. The standard curving would mean that even though these students did much better on the same tests and the same homework would get the same grades as the students from years before. That is simply unfair. They performed better, so they should be rewarded with better grades.

Waves like this occur all the time - why should the good quality classes be harmed and the poor quality classes be given free A's?

As for your 5 point scaling question that isn't exactly what has to happen. A professor could easilly say that each range gets bigger. Suppose you have 80%-100% is an A, 65%-80% is a B, etc. That is scaling, that is fair to everyone, and that will have a reasonable number of people in each grade. But artifically saying only 10% will get an A and that 10% must fail is quite unfair to the students.

In many classes, you get an inverse Bell curve in grades. About 45% get 85% or more, and about 45% get well below a 50%. Basically those who studied did quite well, and those who didn't study did quite poorly. I say in that case give all the top people A's and the bottom people F's. That is fair. But with a curve, can you tell me a fair way of deciding who gets a C? Try these numbers:

My example curve: 10% get A's, 20% get B's, 40% get C's, 20% get D's, and 10% get F's. The grades are:
10 students got a 100%
10 students got a 95%
5 students got a 75%
10 students got a 50%
10 students got a 40%.
Now apply the curve. This is very typical of many classes. I'll check back tomorrow to see if you find a fair way of applying that curve.

 

[Mears]

TallBill, do you want to elaborate on why I am wrong, or do you expect me to take you say at face value?

 

[TallBill]

The issue isnt comparison of universities. Its the comparison of grading. Curves, scales, straight grading.

 

[TNTrulez]

And what are you getting in this class? Your final was today at 8 right? My friend was telling about it. He expects to get a B.

 

[TallBill]

still waitin for someone to grade dullard's class

 

[lukatmyshu]

Originally posted by: TNTrulez
And what are you getting in this class? Your final was today at 8 right? My friend was telling about it. He expects to get a B.

Who were you talking to (my final was at 8 AM this morning and i started this thread, but I haven't posted in in since then).
I think that the curve is designed because every semester we have difference profs, different TAs, all sorts of variables that can change. There is one thing that shouldn't change (usually with a large enough University) ... the INTELLIGENCE of it's students. Since, especially in these courses that are all upper division and in a very difficult major that you have to APPLY to get into after you get into a very difficult university, you assume that the students are always doing a fairly reasonable amount of work. I think that in situations such as the one that I described, it is very fair to say that the mean score is going to be a B. Do better than the mean, and you'll do great in the class. But apply that only if you need to, if the mean is a 95% then give everyone an A ... they deserve it.
If you notice, the 'more acclaimed' institutions not only endorse curves, they live by them. At Stanford the mean is a B-B+, in grad school you either get an A or B. At Harvard Law, 50% of the class graduates with honors. Why? Because at this level, if you can beat 50% of those who got into Harvard Law .... you deserve Honors. I honestly feel that if I can do better than 50% of the students in a CS course at UC Berkeley .... I deserve a B.