April 25, 2004

Foolish and sensible consistencies in grading

The practical problem with reading more than 10 or 15 student exams or papers is that one has to balance the need for consistency with the need for sanity. Even with a good bit of experience and a grading scheme (or rubric, or what-have-you—a set of guidelines for grading), there is always the good chance that one's standards will change from the beginning of a batch to the end. As a colleague tells me, sometimes the paper that you cracked down on for minor factual problems looks a lot better by the bottom of the pile. Conscientious teachers are aware of that and will reread some papers to make sure that the grading standards haven't changed.

But the deeper grading problem is a simpler physical one: reading student work carefully is exhausting. When the exams are handwritten, one can enjoy the neat print-like work of some students but struggle through other work that has overlapping curlicues and daggers of ink or just a muddle that one must puzzle out. Did the student write "curdle" or "coddle"? More fundamentally, you're trying to figure out whether the student understood concepts and can explain them clearly. You don't want students to regurgitate whole phrases (or at least I don't), but a student who puts an idea in her or his own words may also be mangling those ideas beyond recognition. It's a balancing act, and that act tires me out.

The undergraduate exams I'm now grading are all paragraphs in response to eleven questions. So, instead of making a judgment on one or two essays per student, I'm doing so on five or ten times as many. I think it gives me a better sense of what students have learned in this course than long essays, but it's more mental work, believe me.

The practical solution for me has turned out to be partially batch processing, partially blind grading of individual questions. I take a batch of papers (say, 10 or 11) and go through them one or two questions at a time. That theoretically gives me some question-by-question consistency within a batch of papers while still letting me see progress while I work. For example, I have the white-paper exams queued up this morning, having turned the first sheets over last night, so that I'll only know student identities after grading questions 3-11 and then turning to the first sheet and the first two questions. I should finish this batch over four or five hours, but I'll take breaks between questions.

All right: enough procrastinating. Time to start grading this batch. Listen to this article
Posted in Teaching on April 25, 2004 7:17 AM |