Memo 1: Guessing | Memo 2: Difficulty | Memo 3: Essays
Memo 4: Multiple Choice 1 | Memo 5: Multiple Choice 2
Memo 6: Averaging Grades | Memo 7: Assigning Grades
Memo 8: Reliability | Memo 9: Missed Test
Memo 10: Multiple Choice 3 | Memo 11: Absolute/Relative Grading
Robert B. Frary
DEAR ANN: I was a graduating senior in business administration with a B average and seven job offers at a prestigious eastern university. I didn't graduate, the job offers were withdrawn, my parents kicked me out, and now I have to work in McDonalds. All this is because of one zero in a crummy liberal arts elective. Any passing grade would have permitted me to graduate. I got Bs on the first two of three tests and decided to just skip the final, since two Bs and an F average out to a C (3+3+0)/3=2. Right? Not according to Prof. S.! He averaged (84+81+0)/3=55. Since 60 was the lowest D, he gave me an F. His excuse was that he had said he'd average in a zero for anyone who skipped the final. What kind of zero should he have averaged? --PHRANTIC IN PHILLY
DEAR PHRAN: Two kinds of zeroes? I thought there was only one. Maybe you should have written to Marilyn Vos Savant. But, still, I don't think any faculty member at a "prestigious eastern university" would do anything like what you describe. Maybe at a teachers' college, but then I should have noticed the New Haven postmark.
What's going on here? There is a lot of missing information. Did Prof. S. ever tell the class how he would actually average the grades? Why didn't Phran let Prof. S. know of his/her intention to skip the final and confirm how the grades would be averaged? Did the final exam cover all new material? Did Phran attend and participate in class during the latter part of the semester? However, the really fundamental question is how well the grade of F can be defended as a portrayl of Phran's achievement in the course.
It is widely accepted that, barring academic dishonesty, a grade should reflect student achievement. Disregarding Phran's flagrant attitude, it is certainly within the realm of plausibility that his/her overall achievement was not at the F level and that Prof. S. knew this. If that is true, then the F was effectively a punishment. The question of a faculty member's right to punish a student is complex and emotionally charged. At Virginia Tech, it should be noted, faculty members are not permitted to discipline students for acts of academic dishonesty; the Honor System provides for disciplinary actions under conditions designed to ensure due process. Violations of university rules and regulations are handled similarly by the University Judicial System. Punishment for undesirable academic behavior other than that under the jurisdiction of the Honor System is not formally controlled, though a student may appeal the decision of a faculty member through the department head.
Of course, most faculty members would like to avoid such controversies and surely would prefer that confrontations with students over matters such as Phran's complaint not lead to establishing yet another judicial system. But, nevertheless, students do skip tests and fail to turn in assignments. In such cases, the instructor is certainly justified in assuming (conservatively) that work was not done or that material was not learned. More important, the instructor should respond in a way that does not reward delinquency. Consideration of the "two kinds of zeroes" can lead to some workable and justifiable grading policies for tests not taken or assignments not done.
What makes the two kinds of zeroes different is how far each zero is away from whatever number represents the lowest D relative to the total scale range. In the case of averaging letter grades, the range of numbers to be averaged is 0=F through 4=A. A D is represented by the number 1. Thus, an F is one-fourth or 25% of the scale range below a D. Contrast this with a score of zero on a 100 point scale where the lowest D is 60. Then the F is 60% of the total scale range below the lowest D. Obviously the two zeroes do not evaluate achievement in the same way. What should the instructor do?
Some instructors assign letter grades for each test or assignment and average these on a 4-point scale to determine the letter grade for the course. This is what Phran expected Prof. S. to do. Under this system, it is reasonable simply to assign an F for missing work or a skipped test. The achievement for that assignment or test is certainly unsatisfactory and should be averaged with other grades to determine the student's semester grade. This approach can hardly be considered punitive.
Many other instructors record scores for each test or assignment and average these to determine the final grade. In this case, it is necessary to consider the range of scores actually attained by students who are truly worthy of grades of F. If the test is multiple-choice, say, with four options per question, a completely ignorant person is likely to get a score close to 25%. Somewhat lower scores are, of course, possible but can only result from very bad luck or extensive misinformation. Regardless of the testing mode, superficial familiarity with course vocabulary and topics may permit even the weakest or most negligent student to score somewhat above the effective minimum possible score. A reasonable assumption that the instructor can make under these circumstances is that a student who skipped the test would have done slightly less well than the person actually making the lowest F. Assigning this score for missing work also has the advantage of not rewarding a negligent student with a score higher than that of someone who actually did the work.
Some readers may feel uncomfortable with the idea of giving a score of, say, 40% for a test that was not taken. This discomfort probably arises from interpreting test scores in a sort of economic sense: no credit for work not done. In this mode, the test is thought of as representing a set of standards, and the grades the students "earn" are awarded largely independent of any decision making by the instructor, often on a preestablished percentage scale. The idea that a score on a scale of 0%-100% represents the percentage of the knowledge covered by the test that the student learned adds fuel to this interpretation: absence at the time of a test is equated with total ignorance or turning in a blank paper (however improbable that might be).
More plausible and psychometrically justifiable is an approach that views the score on a test as related to but not a direct expression of knowledge. A student with a score of 100% probably does not know everything there is to know about the topic tested, and even a score of 0% does not necessarily imply total ignorance. Given this interpretation, it makes sense, in the case of an unexcused absence, to award the lowest score that any legitimate student who actually took the test could plausibly have made. The goal is to make a conservative estimate of the missing score, not to assign some arbitrary minimum.
Whatever method the instructor adopts for dealing with unexcused absences from tests or failure to turn in assigned work should yield final grades that are realisitic, given the missing information. Granted, the assumption of a low score for missing work may result in a final grade that understates the actual level of a student's achievement. Occasional cases of this outcome are probably unavoidable, given the absence of information and the need to make a conservative assumption concerning the missing achievement information. However, final grades that clearly represent punishment for negligence may well be subject to challenge.
Postscript: Some students reportedly have been under the impression that a university-wide grading scale based on percentage ranges exists at Virginia Tech, e.g., 90%-100% = A, 80%-89% = B, etc. This is untrue, and, furthermore, the imposition of such "standards" is extremely rare in post-secondary institutions in the United States.