I’ve written about the use of peer grading at university before (here and here) and suggested a specific way of integrating it a course (here). After discussing it with colleagues, administrators, and even legal counsel, I’ve thought some more about it. What I want to suggest is still pretty abstract and schematic, but I think the basic idea is sound and I hope to try it out at some point in the future.
Let’s think about this within the framework of a one-semester course, with 10 weeks of instruction (and weekly assignments) and six weeks of independent research, culminating in a final term paper.
- 50% of the grade comes from the term paper.
- 35% of the grade comes from the weekly assignments.
- 15% of the grade comes from how well a student’s peer-grading matches the teacher’s.
Here’s how it would work. Every week the students are required to write a single paragraph about that week’s reading and submit it 24 hours before coming to class.
The teacher reads and grades these paragraphs as part of class preparation. The teacher is given two minutes per paragraph and the students are told that these are the conditions under which the paragraphs are read. They are told to write in such a way that the qualities of their writing and thinking are obvious. The teacher gives the first paragraph 50 points. All the subsequent paragraphs are given points relative to the first. The students are then given a grade A, B, C, D on a normal distribution. (Fs are given only to paragraphs that receive 0 points, which is to say paragraphs that are woefully incomplete.)
Note: In a class of 60 students, needless to say, this will be a lot of work. About two hours. The idea is to count this as a substantial part of the teacher’s prep time. It should be obvious how reading a paragraph from each student about how they understand that week’s reading will be useful in planning and executing the lecture. Learning how to make use of this information may itself be an occasion for pedagogical innovation.
After class, the students are given 5 paragraphs from their fellow students to grade, ranking them from best to worst. The paragraphs are selected randomly but so that each letter grade is represented: there is an A and D in each packet, and at least one B and C. The students are told this in advance.
The students now get a grade according to how well they predicted the teacher’s grade. Getting all five in the right order earns an A. Each mistake costs one letter grade.
I imagine there are many possible objections to this approach. I want to acknowledge, first of all, that this places some pretty strong demands on the teacher, whose grading is now subjected to rather public scrutiny. The students must be told that this is sort of a game and it isn’t always entirely fair. Just as in sports, the umpire can make the wrong call sometimes. The point is that, in the absence of outright corruption, these little injustices average out over the long run. This in itself is an important lesson for students to learn. Quality in writing is not only partly subjective, it is subject to error. Get used to it.
I also want to point out that the students will be asked to assess the quality of the written work of their peers on a weekly basis. This, I want to suggest, is an invaluable experience. They will be asked to notice that it is, in fact, possible to compare written output and evaluate it. They will be asked to detect the differences that distinguish the best work (A) in a sample from the worst (D). All the work may be “good” in its way. They will be asked to decide what made one better than another. They will not be allowed to say that this is meaningless question or an impossible task. Hopefully, this will give them some insight into the teacher’s predicament as an examiner. It may even occasion some empathy.