bigblue

Teaching - Hit Your Mark

Knowing your way around a curve will make grading less arbitrary, and earn you high scores from your students

By P. Wankat & F. Oreovicz

Of all the activities a professor undertakes, grading, though fraught with angst and ambiguity, is among the most important. Students strongly echo this and are adamant that grades must be fair. But are grades an accurate predictor of success?
Grades are ultimately subjective, no matter how much we would like to believe otherwise. For example, using 11 different grading schemes, we calculated the final grades for an actual engineering course. Depending on the chosen scheme, a raw score of, say, 67 could receive anywhere from a B to a D.Hit Your Mark

How Do I Rate Thee...?
With so many grading options, and with so much riding on grades, educators should have a thorough grasp of the different systems so they can make an informed choice. Grading systems all have strengths and weaknesses, but they are not all created equal.

Criterion grading. By setting a predetermined standard for each letter grade, this method reduces competition among students, since one's grade does not depend on the scores of classmates. It also gives students clear goals and a clear idea of where they stand. Indeed, any course that benefits from student cooperation requires some form of criterion grading.

Although criterion grading may seem strictly objective, the arbitrary selection of numerical standards for each letter grade is, in fact, subjective. A popular criterion grading scheme assigns an A for grades of 90 to 100, a B for 80 to 89, and so on; but the scale could be set anywhere.

The combination of test difficulty, which is also subjective, and grading scale sets a test's grading level. With extremely difficult or overly long tests, strict use of criterion grading may be demotivating for students. To compensate for a difficult grading level, you can divide student scores by the maximum score achieved, and then apply the desired scale. This increases all scores by the same percentage. Another method that will reduce student complaints is subtracting the maximum score achieved from 100, and then adding that number of points to each student's score.

Informing students in the syllabus that the scale represents grade guarantees can also soften criterion grading. You can reserve the right to lower the scale, and even make criterion grading quite humane by routinely examining the scores of students who just miss a grade and boosting them to the higher grade when appropriate (a judgment call).

Curves. Because they assign grades by comparing student scores, curves are easily adaptable if test scores are low. But curve grading discourages student cooperation because one high score can "blow the curve" and prevent low scores from being revised upward. This can lead to cutthroat competition, or even deliberate sabotage.

Of course, the curve can be moved up or down at will, so it, too, is subjective. Perhaps the only advantage of curve grading is that one can often reduce student complaints by looking for breakpoints in the scores to use as boundaries.

Professors often use curve grading because test scores are low. It's better practice to write good tests and be sure the students have sufficient time (see "In Praise of Testing," February 1999 PRISM). If a test is a disaster, adjust the scores but continue to use criterion grading.

T-scores. Essentially a variation of curve grading with an aura of statistical objectivity, T-scores convert raw scores into a distribution with a mean of 50 and a standard deviation of 10. Arbitrarily, and subjectively, a T-score of 50 is often set as a C, with a 10-point range for each grade. T-scores have most of the disadvantages of curve grading, and since the scores in many courses do not follow a normal distribution, T-scores are often not statistically valid.

With the exception of the very top and bottom students, every student in your classes could have his or her grade changed under a different grading scheme. Selecting a scheme is not a trivial matter, and deserves as much care and thought as any other aspect of your teaching.

    Phillip Wankat is the Clifton L. Lovell Distinguished Professor of Chemical Engineering at Purdue University. Frank Oreovicz is an education communication specialist at Purdue's chemical engineering school. The authors welcome readers' feedback. You can reach them via e-mail at wankat@ecn.purdue.edu  and oreovicz@ecn.purdue.edu .

    James Stice, a professor emeritus of chemical engineering at the University of Texas-Austin, also contributed to this article. He can be reached at spider@ xmail.utexas.net.



return to PRISM online

bigred