Curving Tests?

sandmangeck · #1 09-27-2011, 02:02 PM

Howdy Folks.
Do you think that curving tests is fair or unfair? I personally think the college students today can't hack it, and have earned the bad grades.

**Eric Perry** · #2 09-27-2011, 02:10 PM

I have never been a fan of curving grades. It's unfair to those who actually did put in the time and effort into studying, and also unfair to those who may be opting to slack off simply "'cause there's a curve". It *will* catch up to them, either in school or in the workforce. Teach them to address their deficiencies now, and hold them accountable for their flippant attitudes!

Jim Nazium · #3 09-27-2011, 02:15 PM

I think that a test needs to indicate whether the students have mastered the assigned material. If tests can't be designed to a uniform level of difficulty, then it's appropriate to make some adjustment based on the students' performance, but a strict curve is a bad idea, IMHO. The incentive becomes just to do average, or better than average, rather than to master the material.

i_got_a_mohawk · #4 09-27-2011, 02:20 PM

I don't see a problem with grading to a curve. Especially with larger sizes.

Standardised high school exams for instance (everyone in the country sits the same exam at the same time and they are marked by independant examiners (not teachers from the school)), I think should have a curve to neutralise the effect of the test being too easy or too hard. And when you have a sample size with tens (possibly hundreds) of thousands, it works.

College/University classes with 100 or fewer people? Nah, no chance. The lecturers should have the exam set at an appropriate level (and, at least where I've studied, that's the way they do it, no curve, just very clear pass/fail points!)

**bassman314** · #5 09-27-2011, 02:36 PM

We asked our college stats prof to curve the grades, before any tests were given.

He said he'd be happy to. Here was his curve:

Median grade: C
+1 Std Dev: B
+2 or more Std Dev: A
-1 Std Dev: D
-2 or more Std Dev: F

So if the Median was say an 85/100, and there was a fair amount of clustering around an 85, you could theoretically get an A at 88 and fail at 82.

No one was willing to risk that.

One of the prof's I had for most of my actual degree classes "curved" as follows:
Highest grade = 100/100
Adjust all others by adding the difference between the actual score and 100 to all other scores.

So if the highest grade on a test was a 95, everyone else got an extra 5 points. Then he used a standard grading scale (90+ is an A, 80+ is a B, 70+ is a C, 60+ is a D, 59 and below is F)

The only time this was an issue was on a particularly difficult exam where the highest grade was a 72.

**Eric Perry** · #6 09-27-2011, 02:42 PM

Quote:

Originally Posted by i_got_a_mohawk

...College/University classes with 100 or fewer people?...

Damn! The biggest class I've ever had (in grad or undergrad) was around 40 people. I couldn't imagine a lecture hall filled with 100+ people! My class size average all through my college years was probably around 12-14 people. I have had a few classes (grad classes) with only 4 or 5 students.

Frank Tuesday · #7 09-27-2011, 02:55 PM

Grading on the curve? Or adjust the maximum number of points based on the top or some other score?

It depends on how you perceive the grading scale. Is it a method of ranking students, or a measure of mastery of subject matter.

If you are trying to rank students, the curve makes perfect sense. If everyone scores above 90%, the ones that score 100% are clearly performing higher than the ones that score a 90%. If the test scores range for 8% to 43%, the relative performance of the students is reflected quite clearly.

If you are trying to measure competency, the curve is not appropriate. This is what you should know, you either do or you don't. This, of course, assumes that the test was fair, and an accurate representation of the material that was presented and the students were expected to know.

Sliding scale (adjusting maximum points) neither ranks the students, nor measures competency. Again, this is for a "fair" test only. For an "unfair" test, one in which the time allotted is insufficient or the test covers material to which the students were not exposed, the sliding scale can be applied to give a reasonably accurate measure of competency.

iamlowsound · #8 09-27-2011, 03:09 PM

If the test is constructed properly, it should naturally curve itself. I have never had a test graded on a curve, which I am glad for.

lowsound

jp58 · #9 09-27-2011, 03:20 PM

Most professors curve tests here because they have to use standardized tests designed by the department as a whole (or just the head of the department). Guess what happens when you get a fluoro-organic professor teaching gen chem lectures. . . Needless to say he didn't cover a lot that was on the tests and over covered other parts. So you have half the class making in the 50 range, with the best students making an 80. I feel like in situations like that, the professor is responsible for covering the material needed or the questions that were not covered need to be thrown out.

i_got_a_mohawk · #10 09-27-2011, 04:32 PM

Quote:

Originally Posted by Eric Perry

Damn! The biggest class I've ever had (in grad or undergrad) was around 40 people. I couldn't imagine a lecture hall filled with 100+ people! My class size average all through my college years was probably around 12-14 people. I have had a few classes (grad classes) with only 4 or 5 students.

Some of our 1st year lectures were generic for the entire school, so had ~100-200 students. 2nd year onwards it was generally 40 or less. Down to about 10 or so by the final year. (19 students graduated within my general field (3-4 different degree types) ).

Most of our grad lectures only had 4 people in them too.

ADbassman · #11 09-27-2011, 04:33 PM

Yes I think curving is fair when the tests are flawed. For example, the course of instruction I'm in right now is considered a pilot program, so a lot of times the tests are completely new and are often put together a week before we take them. The testing department does not talk to the curriculum department, so we have things that show up on tests that were never taught nor within the scope of what we were studying. Truely oddball questions. If they didn't curve the tests, on average at least 75% of the class would fail, and we had to do a lot to get here in the first place so I'm not convinced it's the students.

**jmattbassplaya** · #12 09-27-2011, 08:29 PM

Quote:

Originally Posted by sandmangeck

Howdy Folks.
Do you think that curving tests is fair or unfair? I personally think the college students today can't hack it, and have earned the bad grades.

Grades are almost never curved, anyways. There's always some ***hat who makes an A and ruins any chance of a curve.

**jmattbassplaya** · #13 09-27-2011, 08:30 PM

Quote:

Originally Posted by i_got_a_mohawk

Some of our 1st year lectures were generic for the entire school, so had ~100-200 students. 2nd year onwards it was generally 40 or less. Down to about 10 or so by the final year. (19 students graduated within my general field (3-4 different degree types) ).

Most of our grad lectures only had 4 people in them too.

I'm a 4th year and still have some classes with over 500 students...

**Eric Perry** · #14 09-27-2011, 08:42 PM

Quote:

Originally Posted by jmattbassplaya

I'm a 4th year and still have some classes with over 500 students...

sandmangeck · #15 09-27-2011, 09:05 PM

I posted because the lady and I go back and forth, because the lady will come home with bad test scores and still get b's. I told her I was going to go back for my doctorate if it's going to be that easy

i_got_a_mohawk · #16 09-28-2011, 04:21 AM

Quote:

Originally Posted by jmattbassplaya

I'm a 4th year and still have some classes with over 500 students...

Quote:

Originally Posted by sandmangeck

I posted because the lady and I go back and forth, because the lady will come home with bad test scores and still get b's. I told her I was going to go back for my doctorate if it's going to be that easy

No such luck with that over here, bugger all taught/tested element!

hrodbert696 · #17 09-28-2011, 04:58 AM

I never give tests, I give papers, and no curving. Everyone does A work, everyone gets an A: everyone does C work, everyone gets a C.

PSPookie · #18 09-28-2011, 05:50 AM

Quote:

Originally Posted by sandmangeck

Howdy Folks.
Do you think that curving tests is fair or unfair? I personally think the college students today can't hack it, and have earned the bad grades.

Where I went to school (GT), it was not uncommon for the class average on a test to be in the 40's.

Yay, curve!

Chebass88 · #19 09-28-2011, 06:36 AM

I've had few classes where the grades were clearly spelled out. (90-100 = A, 80-89 = B, etc.) For the majority of my classes, I know the relative weightings of exams (e.g. exams 70% of grade, homework 15%, project 15%), but how the numerical scores translate into letter grades is not disclosed.

A professor SHOULD be able to write an exam properly so there is a natural distribution within the class. I just took an exam yesterday - 6 questions, ranging from very easy to moderately difficult, worth anywhere from 10-20 points. If you paid attention in class & did the homeworks, you could easily get 50%, and then work a little bit for the remainder.

My best classes were with established grading scales and "no curve" professors. Your earned score meets muster or it does not. That is a great feeling - you get an exam back and know exactly where you stand.