# How do I grade on a curve?

Discussion in 'General Education' started by heavens54, Dec 30, 2014.

1. ### heavens54Connoisseur

Joined:
Feb 1, 2007
Messages:
1,785
0

Dec 30, 2014

I've never done it. We had a math test just before break that we had to give. They weren't ready, but I figured Wednesday was the latest I could go. They weren't focused at all. Many new skills thrown at them, dumb pacing guides. I want to give them some wiggle room. What is the best way to do this? My math top kids missed only one. The next best grade is a low 80. The rest are disaster. I want to revisit the skills, but end of quarter is coming up soon. Any guidance on how to do this? Thank you.

3. ### Linguist92021Phenom

Joined:
Feb 4, 2010
Messages:
4,774
999

Dec 30, 2014

I've only done it once so I might be explaining wrong, but the way I remember it is this:
take the top grades, for example you had one kid with 95 % (ignore that one because it's just one student), and you had a few with 89% as the top. Now that 89 % becomes your 100 %, so 89 % is A, 79 % is B, 69 % is C, and so on.

4. ### FourSquareFanatic

Joined:
Aug 4, 2009
Messages:
2,837
315

Dec 30, 2014

That weird outlier kid gets an A and then everything else is out of what the low 80 was. This only works if everything's weighted the same, I guess. Ex: If a few are at 16/20, then everything is out of 16. 16=A.

5. ### swansong1Virtuoso

Joined:
May 19, 2007
Messages:
6,535
1,448

Dec 30, 2014

I'm not that familiar with grading on a curve. Wouldn't that artificially inflate grades?

6. ### msleepRookie

Joined:
Nov 1, 2011
Messages:
91
34

Dec 30, 2014

grading on a curve is grade fraud. It does not represent what they know or do not know.

You should have them take it again.

7. ### mathteachertobeCohort

Joined:
Jun 20, 2011
Messages:
593
119

Dec 30, 2014

If the test was assessing skills and concepts your students hadn't learned yet, then I think the greater fraud would be to let the grades stand as is. Is there any way you can use the test as a pre-assessment? The school or district may have required it to be given, but is it also required the grades go into your grade book?

As far as curving goes, look at the distribution of grades. If it resembles a normal curve at all, then you might allocate 10% As, 20% Bs, 40% Cs, 20% Ds and 10%Fs. I've taken many classes where the letter grades were distributed based on class averages and standard deviations. It is neither fraudulent nor inflationary, it simply means the test was not written with the expectation that well-prepared and capable students would be able to answer all of the questions.

Joined:
Jun 14, 2013
Messages:
3,877
1,485

Dec 30, 2014

I don't think grading on a curve is grade fraud. It was a common practice when I was in high school and college. I get why some don't like it, but there is a lot in education that is debatable. I haven't heard of it in elementary school, though.

9. ### 2ndTimeAroundPhenom

Joined:
Aug 10, 2010
Messages:
4,330
572

Dec 30, 2014

It is grade fraud. It is giving points that students did not earn.

If you think the kids got a raw deal, either retest or throw the grades out all together.

10. ### Linguist92021Phenom

Joined:
Feb 4, 2010
Messages:
4,774
999

Dec 30, 2014

My students in general are much lower than your average kids; a lot have been locked up and spent half of their time with watered down / dumbed down curriculum and assessment, others have ditched school so much, or were always kicked out of class that they're naturally lower. The higher kids, who seem to stand out from the crowd as the super bright kids are your average smart kids in 'regular' school. I'm not saying my kids are not smart, but they have a disadvantage.

Yet, my expectations are much higher, for me an A is truly an A, and 59 % is still an F. I would say the only thing I grade nicer on is writing essays.

I have used curved grading only one time and it was to the suggestions of my P. We were doing poetry and spent about 3 weeks on it (as much as my students can stand) gave them a final assessment and was ready to move on. I don't know what happened, but almost everyone did worse than I accepted. Even my higher students, or those who were obviously getting poetry just didn't do well. I didn't want to fail half of my students, but I didn't think reteaching all that would be beneficial. My P suggested curving and I think she was right.

I don't think I would use this a lot, this was last year and have no intentions of doing it again.

11. ### Reality CheckHabitué

Joined:
May 29, 2010
Messages:
912
29

Dec 30, 2014

Take all of the scores and list them highest to lowest.
Then:

Top 2 - A's
Next 3 - B's
Mid-section (largest) - C's
#5, #4, and #3 from the bottom - D's
Lowest 2 - F's

12. ### 2ndTimeAroundPhenom

Joined:
Aug 10, 2010
Messages:
4,330
572

Dec 30, 2014

While this will never be tolerated in a public high school, it is probably the best, most genuine type of curve that exists. While it doesn't truly assess the mastery of standards ( no curve does) it does at least determine who has more mastery and who has done the best with the information given.

13. ### Reality CheckHabitué

Joined:
May 29, 2010
Messages:
912
29

Dec 30, 2014

I've been in public school most of my career and it's not only tolerated, it's strongly encouraged and embraced with enthusiasm.

14. ### heavens54Connoisseur

Joined:
Feb 1, 2007
Messages:
1,785
0

Dec 30, 2014

On some of the skills, the curriculum we used wasn't very effective in teaching the skill. They were distracted with Christmas activities and other things going on that last week. It was the hardest test they've had. I just want to help them out. On a pacing guide, we move on to a new chapter. I could take out some questions, but I didnt want to spend a lot of time on it.

15. ### cartwheelsRookie

Joined:
Jan 24, 2011
Messages:
65
0

Dec 30, 2014

I had been grading on a curve this year at my now-former school, but only for one specific thing. The school is very big on open-ended responses, and they had a mandated rubric that we were required to use, and it held the students to INCREDIBLY high standards. Now, I taught self-contained, and I still had to use this rubric. While I was teaching to the rubric and pushing my students, it simply would not have been fair to give them the raw score. Therefore, it would look something like this:

9/15 is the highest scores, so it becomes 100% (15/15), so I would add 6 points to every student's score (5/15 is now 11/15, 2/15 is 8/15 and so on).

This was the only place I used a curve, though, as not curving would be holding my students to an absurd standard (even some gen-ed teachers curved their grades on this rubric).

16. ### a2zMaven

Joined:
Sep 16, 2010
Messages:
5,820
1,750

Dec 30, 2014

The problem with grading on a curve, especially the bell curve, is that you aren't comparing the students to the assessment of the material taught but to how well the other students in the class learned the material (or classes if the curve is used across multiple classes). Since we are supposed to be assessing students based on what they are supposed to learn and be able to demonstrate via the assessment, the bell curve does not give an accurate picture of how well the students learned the material which is what grades are supposed to represent.

17. ### olivecoffeeCompanion

Joined:
Oct 26, 2013
Messages:
188
0

Dec 31, 2014

My 3rd grade cooperating teacher would cancel out questions if 10 or more students missed them. So, if the test was worth 20 points and 10 students missed Q13, Q13 would be cancelled and the test would then be worth 19 points. I've seen her cancel out as many as 5 questions, reducing the total value of the test to 15.

18. ### 3SonsEnthusiast

Joined:
Aug 30, 2007
Messages:
2,008
169

Dec 31, 2014

I don't agree that grading on a curve is grade fraud. There is a danger of misrepresenting knowledge, but a good teacher shouldn't be using only one measure anyway. A graded test (curved or not) should be just one component of an overall grade.

Reality Check probably has an okay rule of thumb, at least for the right class size. I believe there's really a lot more math to it than that, including finding the mean and standard deviation. Otherwise his technique could lead to some unfairness if strictly applied (imagine scores of 87, 87, 86, 64, 62, 52, .... 32, 31, 30 -- given those kind of scores, it wouldn't really be accurate to give the one who scored an 86 a B, lumping them in with the ones who scored in the mid-sixties.)

19. ### Linguist92021Phenom

Joined:
Feb 4, 2010
Messages:
4,774
999

Dec 31, 2014

That makes more sense than grading on a curve, actually. If I had a lot of students miss certain questions, that would be an indicator to me that I probably didn't do a good enough job teaching those concepts. Canceling them out is fair, and I can reteach those.

Joined:
Jul 19, 2014
Messages:
10,381
2,565

Dec 31, 2014

21. ### a2zMaven

Joined:
Sep 16, 2010
Messages:
5,820
1,750

Dec 31, 2014

Isn't fraud misrepresenting something on purpose?

If there are multiple teachers teaching a subject and one teacher curves because the students didn't learn the material, isn't that academic fraud for whatever reason the student learn the material?

There are ways that students demonstrate skills, but not all are equal in what is demonstrated. Group work that demonstrates skills is not independent work. That can't be used to demonstrate independent mastery. Demonstrating during classwork immediately following instruction can't demonstrate independent longer term understanding and mastery which is really what is necessary for students to have.

So, curving is a form of academic fraud, particularly if the test isn't designed to have a projected top score below 100% which I have seen done in classes before. But at that point, the teacher or professor knows that it would be remarkable for a student to achieve a perfect score. The teacher or professor already has a known mark of A level achievement. Curving because a teacher didn't get desired results is very different.

22. ### heavens54Connoisseur

Joined:
Feb 1, 2007
Messages:
1,785
0

Dec 31, 2014

Thanks all. Maybe grading on a curve is not what I'm looking for in this situation. Maybe throwing out the questions that I knew they would struggle with, and did, is maybe the better answer. Thanks for your suggestions. It's frustrating sometimes to find the right solution for my students.

23. ### 2ndTimeAroundPhenom

Joined:
Aug 10, 2010
Messages:
4,330
572

Dec 31, 2014

That is not grading on a curve, though. I have no problem with throwing out bad questions, especially if the standards are still met with another assessment (or other questions on the test). If they aren't, the teacher should reteach and then reassess. That's just good practice.

24. ### heavens54Connoisseur

Joined:
Feb 1, 2007
Messages:
1,785
0

Jan 2, 2015

OK, I analyzed my grades and this is what I have on that test;

A's - 2
B - 1
C - 5
D - 4
F - 6

Does this reflect a Bell Curve? I guess I don't really understand the curve philosophy.

On one question, 16/19 missed. On two questions, 12/19 missed. Would you throw out all three of those questions for all tests? Or just the ones who missed? Grrrr. Just not sure what to do. I do know that this test had too many skills, and some very advanced. our curriculum did not provide enough support for them to succeed on the test. It was two days before Christmas break.

What would you do?

25. ### 2ndTimeAroundPhenom

Joined:
Aug 10, 2010
Messages:
4,330
572

Jan 2, 2015

You keep saying that it was two days before Christmas break. I don't think this should matter at all. I gave tests the day before Christmas break. Granted, I teach older children, but if you have the "all business" attitude for those days before break, it shouldn't make a difference. I definitely DON'T think the kids should be cut a break on their test scores because of the date. That is precisely why *I have problems in high school with students thinking that the whole week before Christmas "shouldn't matter."

I would revisit the material, especially the questions that were missed by most of the students. I'd review the test with the whole class. This would take me 30 minutes max. The next day we would take a retest. With this age group I would take the higher grade of the two. With my age group I do a variety of things, really depending upon how much I feel the students prepared on their own for the first test.

If you have more than 30 minutes allotted for that subject on your first day back, I would review for 30 minutes and then introduce new material. I'd give them the evening (or maybe two) to study - I would not hold up the new material nor would I test again on Monday after the review.

I'd plan for the new test to take less than 30 minutes so my total time dedicated to the retake would only be about an hour.

26. ### kan3288Rookie

Joined:
Jan 3, 2015
Messages:
5
0

Jan 3, 2015

I've used several different techniques when tests turn out worse than I expected.

1) If almost all the students missed a question(s), I would drop that from the total score
2) If there is a question(s) that you believe was unfair or felt you should have gone over more, you can drop that from the total score
3) If there wasn't a bank of questions that were missed or unfair, then your only two other options are to reteach and retest or curve it

I've curved using several methods. I have created my own bell curve before (as someone previously suggested) where I tally up each percentage and determine the "cut-off" percentages for A, B, C, D, and F.

I've also "curved" by allowing students to take a retest after a couple days. It is optional so those who did decently may not want to take it. I will usually go over some of the material again in class, and I expect them to come in individually to get help before they retest. Then, I average the grade from that test with the original test. Typically, if they do worse on the retake I don't change the original grade.

Finally, there is a formula curve I learned from a fellow teacher. He takes the square root of their percentage and then multiplies by 10 to get the new percentage (i.e.- a student who got a 60%, you would take square root (60) then multiply by 10). I think this is somewhat of a generous curve, so I typically don't use it, but it is an option. The one good thing about this method is that it tends to bring up the lower grades much more than the higher grades. So, if you have some students who got A's and B's they receive less of a curve than those who got D's and F's.

27. ### 2ndTimeAroundPhenom

Joined:
Aug 10, 2010
Messages:
4,330
572

Jan 3, 2015

Finally, there is a formula curve I learned from a fellow teacher. He takes the square root of their percentage and then multiplies by 10 to get the new percentage (i.e.- a student who got a 60%, you would take square root (60) then multiply by 10). I think this is somewhat of a generous curve, so I typically don't use it, but it is an option. The one good thing about this method is that it tends to bring up the lower grades much more than the higher grades. So, if you have some students who got A's and B's they receive less of a curve than those who got D's and F's. >>

This is similar to what a lot of my fellow coworkers do. I hate it. It isn't fair to the students that did their work but missed more than they would have liked. It rewards the students that do poorly with a bigger curve. Of course, almost everyone I work with is a die-hard liberal that thinks everyone should have the same job, standard of living, number of trophies, grades, etc., so it shouldn't surprise me.