Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: [Phys-l] assigning grades



Dan Crowe already made the major point that I would like to make. I'll expand it a bit more.

I believe the grading system that works the best is a final percentage that is the weighted average of the percentages on various kinds of work.

In my classes the exam average is weighted 3, lab average is weighted 2, and problem-set average is weighted 1. That means the exams are 50% of the final grade, the labs are 33%, and the problems are 17%. I have sometimes played with the weights, but this is what I am using right now.

With this system, as soon as you have one exam, a lab report, and a couple problem sets, the student can easily see where their grade is headed, and which area might need improvement. In a total-points system, as Dan said, the problem set points and lab report points grow (between exams) and make it appear the student is doing better and better, then when the big-point exam comes, and the student does less well, the grade suddenly drops.

In the weighted percentage system suppose the student gets 70% on the first exam, has a 90% average on the first couple labs, and a 95% average on the first several problem sets. According to my weights, the total weighted average is about 81%. Between the first and second exam, if the student continues to turn in problem sets that are running 95% and lab reports that are running 90%, then her grade doesn't change one bit during that time. By contrast in a total point system it will appear the grade is increasing, then if the student gets another 70% on the next exam, the grade will drop back to what it was after the first exam.

It's possible to have a combination of total points and weighted percentages, and I do this most notably with problem sets. Some of my problem sets have 3 problems and some have 8. It might make sense for the 8-problem set to count more. Thus, I record "points out of possible points" for each problem set. The student has to keep a tally of their total points and the total possible points in order for them to figure their problem set percentage. But no matter how many problem sets and problem-set points there are, the overall percentage on the problem sets will still represent 1/6 of the final grade. I don't have to worry about how many problem sets I have, nor how many points are on each set. I just assign problems as I see fit, and some of them take longer and some don't.

On the other hand, all my labs are the same value, so I can just record the a percentage grade for each lab report. All my exams are also worth the same. I don't want any one exam to count more than the others. So I just record the percentage for each exam.

This system makes it very easy for the students to see where they are, and it is easy to apply. It's also somewhat sad that they don't change much throughout the term. Students tend to be pretty consistent on their problem sets, and these run high because they can get assistance. After the first lab report, where some of them learn that they need to follow directions, the lab grades tend to settle into a consistent percentage. And the by the second exam you pretty much know where the 3rd and 4th exam are going to be. The best way for them to improve their grade is to do better on the exams, but they seldom do.


Michael D. Edmiston, Ph.D.
Professor of Physics and Chemistry
Bluffton University
Bluffton, OH 45817
(419)-358-3270
edmiston@bluffton.edu