I use percentages to map to letter grades.The percentages mirror the 4.0 scale, except that where a GPA difference of 1.0 corresponds to a full letter grade,I use a percentage difference of 10. The table below shows the conversion from numerical grades to letter grades.
Number → Letter Conversion | |
|---|---|
Numerical Grade | Letter Grade |
≥ 97.5 | A+ |
≥ 92.5 | A |
≥ 90.0 | A- |
≥ 87.5 | B+ |
≥ 82.5 | B |
≥ 80.0 | B- |
≥ 77.5 | C+ |
≥ 72.5 | C |
≥ 70.0 | C- |
≥ 67.5 | D+ |
≥ 62.5 | D |
≥ 60.0 | D- |
< 60.0 | E |
Often I will grade with letter grades on subparts of an assignment, convert to numbers for averaging (or weighted averaging), and then back into letter grades using the above table. For conversion from letter grades to numerical grades, I use the middle of the numerical range above. Thus, an A is a 95, halfway between 90 and 100. An A- is a 91.25, halfway between 90 and 92.5. Etc.Grades between these are averages. For example, an A/A- is numerically (95+91.25)/2=93.125, which is an A slightly lower than 95/A.
A frequent complaint is that you "lose 5 points" for a 95=A. But I treat 100 as more like an A++. So the few students who achieved a perfect score in all 7 quizzes deserve 100.
Here is the letter → number conversion more precisely:
Letter → Number Conversion | |
|---|---|
Letter Grade | Numerical Grade |
A+ | 98.75 |
A | 95.00 |
A- | 91.25 |
B+ | 88.75 |
B | 85.00 |
B- | 81.25 |
C+ | 78.75 |
C | 75.00 |
C- | 71.25 |
D+ | 68.75 |
D | 65.00 |
D- | 61.25 |
E | 55.00 |
Why I don't round grades
It is my practice not to round the numerical grade before mapping to letter grades by the table. This can be a sore point, so let me explain. For example, I use ≥90.00 as the transition from a B+ to an A-. This means that if your numerical grade is 89.9, I map it to a B+ and not an A-. It can be heartbreaking to miss a grade boundary by -0.1,—or even by -0.01—I know! This is exacerbated by me mostly grading with A's and B's, with C's and D's rare, so that many grades hover around the 90.00 boundary.
But to round up, say, every numerical grade ≥89.50 to 90.00 and map that to an A-, means that the transition from B+ to A- is actually 89.50, not 90.00. And that would mean that a grade of 89.4 would miss a grade boundary by -0.1. (It would also mean that me announcing the grade boundary of 90.00 is inaccurate.) No matter what policy is followed, some could miss a grade boundary by a hair. Even though there may be some psychological difference between the two situations, I prefer to keep it straightforward by announcing the sharp grade boundary from day one and then following it strictly. I find it helps keeps the process more objective, and does not allow room for subjective grade adjustments, which are almost always unfair.
