The independent news organization of Duke University

AI imparts unfair academic conditions

The similarity of the new grade averaging system and computer football rankings has not been noted, but has critical implications. The latter take into account both league standing and the relative strength of the league. So too must the grade averaging.

If standing in a particular course is emphasized, then the bottom student in a high-quality 200 course of five students with strong interest and background is in the bottom fifth of the class just like one in the bottom fifth of a large lecture course. The bottom team in the SEC would be treated like the bottom team in the Ivy League. Surely that is inequitable, and surely the scientists do not want the bottom students in a 200 chem course to be treated the same as those at the bottom of Chem 20.

For this reason, "quality" of the students in the class must be a key element in the averaging formula and perhaps the dominant element. Quality is measured by the achievement score, and that can be cumulative. If the "A" students in Chem 20 become chemistry majors, then presumably the first 100-level chemistry course they take will be weighted very highly. And when they move on to a second 100 course, it too will be weighted highly, etc.

Those who were in the bottom quarter of the beginning science classes may decide that science is not for them and move in other directions. How much will their first poor grades in science pull down the achievement scores in the courses and majors to which they move? Do those courses become Division II so far as student class standing is concerned? This seems quite possible, all the more so because beginning math and science do not seem to be graded on the current University scale, but to have very low grades for the bottom of the class.

The really explosive question involves minorities, especially those who were admitted with lesser records for affirmative action or athletic reasons, as well as foreign students in majors in which the knowledge of English is important. If they do poorly their first year, this will affect their achievement scores in later courses unless they thereafter take courses in a random manner.

If, for example, black affirmative action students take a year to adjust to the University, but tend to take black history in their second year, that could seriously devalue all grades-including that of the top student-in the black history course and thereafter in similar courses.

The University lawyers need to think extremely long and hard on this question. Surely the various sunshine laws would require that all students know the average achievement score in any course in which they enroll and have the right to withdraw if it seems disadvantageous. And if the average score changes after the first day, they need to know that and again have the option of withdrawing.

Surely a new grading system that ensures that the best blacks and/or other minorities taking multicultural courses or majors systematically have a lower "achievement" class ranking than they would a traditional class average is open to serious and successful legal challenge.

Of course, it all depends on what formula is adopted, and a wide variety of results could be achieved. But this is the biggest problem of all: The process of averaging is not intuitively transparent. The losers inevitably will suspect the formula is being manipulated to serve some other sub-group in the student body. All this would be spread over the national newspapers with students of one set of interests or another being told that the University is not a hospitable place to enroll.

Jerry Hough

James B. Duke Professor

Department of Political Science


Share and discuss “AI imparts unfair academic conditions” on social media.