Some Questions and Answers about Grades

These are the personal views of Hamilton Richards, and may not be entirely consistent with those of the University of Texas or of its Department of Computer Sciences.

Are grades necessary?

In principle, no. At some colleges and universities, grades are dispensed with in favor of narrative evaluations. The assumption is that instructors get to know their students well enough, and have time enough, to write a page or two for each student, describing in depth the strengths and weaknesses of the student's performance in the class.

Narrative evaluations are clearly better than simple letter grades, but they are possible only in small classes --say, 20 students at the outside. And small classes mean low student/faculty ratios, which usually translates into high tuition (typically >$20K per year) or large state subsidies.

At UT, where tuition is low and state support is meager, classes are much too large for narrative evaluations, and letter grades are standard.

What purpose do grades serve?

Grades are intended to reflect the degree to which students have mastered the course material. Although we commonly say that instructors "give" grades to students, we don't mean that grades are gifts from the instructor; they are earned by the students who receive them.

Why does UT use letter grades instead of numbers?

Good question. It seems a pity to evaluate a semester's work by choosing one of only five grades (A,B,C,D,F). In large classes, it means that two students whose semester scores differ by one percentage point can get different grades, while in the same class two students whose scores differ by nine points can get the same grade.

It would be much better if we could turn in numerical grades, or at least qualify letter grades with "+" or "-".

Are individual test grades "curved"?

As the term "curve" is generally used in connection with grading, it refers to a mapping from numerical scores to letter grades. When people ask whether grades are "curved", they usually seem to be asking whether the mapping is something other than the conventional 90 -> A, 80 -> B, etc.

I've never seen any reason to assign letter grades to individual tests. Letter grades obliterate information, and they introduce boundaries that are completely artificial. Assigning letter grades makes 79 indistinguishable from 70 but essentially distinct from 80. Letter grades are indefensible; they persist because of a combination of institutional inertia and a certain appeal to the numerically challenged.

I assign letter grades at the end of the semester, because that's what the university requires. Since I don't assign letter grades at any other time, the question of whether there is a curve on an individual test is meaningless.

I worked really hard this semester. Don't I deserve a higher grade?

Grades are supposed to reflect students' output, not their input. A grade should indicate what a student has learned, or is able to do, at the end of the course.

Consider two different students in the same class. One is very quick, and is able to master the material with very little effort. The other makes up for his lack of brilliance with hard work, and earns the same scores on all the tests as his more brilliant classmate. Would you penalize the brilliant student for having worked less hard?

Suppose we were to decide, nevertheless, that grades should depend partly on students' effort. Can you suggest how that effort could be measured?

Homework takes me many hours a week. Why does it count only 10%?

When we say that students earn grades, we should not be misled into thinking of grades as a form of hourly wages. They are supposed to reflect the quality of the student's work, rather than its quantity.

A class like this one presents both opportunities for learning and opportunities for demonstrating what one has learned. Homework assignments are opportunities for learning; since we have no way of knowing for sure who actually did the work, they are not useful for measuring students' mastery of the material.

The D I received is the highest D in the class, and I need a C to move on to the next course. Can't you raise my grade to a C?

You are of course welcome to check your final exam, to make sure it was graded correctly. If there is any mistake in the grading, or in computing your overall average for the course (according to the formula published in the course syllabus), I will gladly correct it.

The notion of grading error must, however, be defined rather narrowly. For example, I am not willing to alter the number of points that are deducted for each kind of mistake. Students, because they are interested parties, are not in a position to offer useful advice on that subject.

You should recall, moreover, that your homework performance has already been included in the computation of your course average, and that by the time final grades have been issued, the deadlines for reconsideration of midterm-exam and homework grades have long since passed.

Finally, barring grading errors on the final, there is no way your course grade can be changed. Once I have set the letter-grade cutoffs, they are fixed. Setting these boundaries is a responsibility I do not enjoy, but it is obvious to me that sharing it with students would not make it any easier.

I had serious non-academic difficulties this semester. Can't you take them into account when assigning my grade?

Let me remind you of something you already know: A student's grade in a course is supposed to reflect the student's performance in that course.

If an instructor ever starts letting students' grades be influenced by their accounts of how hard they've worked, or of the non-academic difficulties they've encountered during the semester, the word gets around that grades can be raised by means other than performance. Some students choose that route, and the hard-luck stories come pouring in. Now the instructor's problem is to determine which of the accounts are genuine, and then to evaluate their worth, in terms of grades. I don't know of any way to do this fairly, so I don't do it. At all. Ever.

The university supports this policy by specifically prohibiting grade changes except where an error has been made, or to replace an X (which indicates a temporary delay in reporting the final course grade) with a final grade. And the rules about X grades specifically prohibit using them to allow students "time to prepare coursework in addition to that assigned to the entire class" [General Information 2003-2004].

If you believe that your problems during the semester prevented you from studying effectively, you might try taking your case to the dean's office. The advisors there have the authority to grant a retroactive Q or W (they do so rarely). That would remove the D from your record, but you would still have to repeat the course in order to receive credit for it.

I'm taking this course for the second time. If I fail it I won't be able to take it again, and it's reqiired for my degree program. So I simply have to get at least a C.

A student for whom the consequences of failing the course would be especially severe should be most careful to stay safely above the dangerous C-D borderline. Students who ration their effort too closely sometimes miscalculate, and are dismayed to find that grades are awarded on the basis of their performance alone, and are not affected by the grades' possible consequences.

Parting thoughts

If you've read this far, you may be feeling a bit put off by so many words about what amounts to a necessary evil. Rest assured that the vast majority of students never have occasion to ask any of the questions concerning poor grades-- they concentrate on getting as much out of each course as it has to offer, and the grades take care of themselves.

So I hope you'll take grading issues in stride, and not let them get in the way of an education that's challenging, empowering, enlightening, --and fun!


Comments and questions to .

This page was last modified on Wed, 20 Aug 2003.