Thursday, April 27, 2006
Grading: It's What They Pay Us For
I’ve heard that a zillion times, and it’s the one part of the job I truly dislike. And soon it will be the only thing keeping me from blasting “School’s Out” by Alice Cooper. Yet it’s vitally important to students and we have a serious professional obligation to do it well. So, what makes for good grading? Here’s my two cents.
First, you must start with a good, fair exam. Err on the side of being too hard rather than too easy, because you can always decide that 70% right is an “A” and 60% right is a B, but if everyone gets practically everything right, it’s hard to make valid distinctions. Of course an exam can be too hard, but the most common mistake for beginning folks along those lines is not asking questions that nobody can answer, but rather asking too many questions/having too many issues. In my first year, in one exam, not a single student finished the last essay: that essay wasn’t particularly hard, nor were any of the others; there were just too many of them. Also, spell out jurisdictional issues in the instructions. In my labor and employment classes, I say in the instructions, “assume all the employers/employees/unions are covered by [the relevant laws] unless the facts indicate otherwise.”
Put in issues of varying difficulty: relatively easy ones so that the “C” papers can be distinguished from the “F” papers, and relatively hard ones, so that the “A” papers can be distinguished from the “Cs.” While you can’t test everything, keep some sense of proportion between how much it counts on the exam, and how much time you spent on it in the reading and in class. Mix in questions that actually have right answers (plaintiff would win here, for sure, and here’s why), with questions that don’t have obviously right answers that make students articulate and weigh competing rules/policy concerns/types of arguments.
Do a detailed grading key. Figure out how many points you should give for certain observations. Then be consistent. If the right answer is “Plaintiff wins, because the one-bite rule doesn’t apply to pet alligators,” four points are available, and a student says, “Defendant wins, because the one-bite rule does apply to alligators,” decide whether that’s worth 0, 1, or 2 points and be consistent for all similar answers.
People vary, but I don’t read Exam 1, then Exam 2, then Exam 3, etc. I read Exam I, Essay A, then Exam 2, Essay A, then Exam 3, Essay A, etc. This helps with consistency.
Read carefully. There is no substitute for that. Make comments, especially on the ones that will get lower grades, so you can remember, explain, and possibly defend what you did.
Essay questions often have kinks. If a question’s wording really was ambiguous or legitimately suggested an issue you hadn’t thought of before, give students credit if they have a plausible argument. I once used the phrase “worked in a large women’s clothing store” in an employment discrimination exam. I meant “large” to refer to the size of the establishment to indicate coverage by a statute. Some students, however, interpreted that to mean “clothing store for large women.” That “fact” arguably could have gone to a legal issue in the problem (for reasons I won’t bore you with). So ... they got credit.
Worst case scenario: some part of the question really makes no sense or contradicts itself in some important way that made it unfairly confusing. Consider seriously just not counting that portion of the exam.
Finally, when reviewing exams with students, I tell them that I will only change a grade if (a) there is a math mistake in calculating it; or (b) if it appears I actually didn’t read or completely misread a section of their exam and didn’t give them any credit when they deserved at least some. Reviews are about the student learning how to do better, not for arguing that knowing that the one-bite rule applied should really be worth 3 points, not 2.
TrackBack URL for this entry:
Listed below are links to weblogs that reference Grading: It's What They Pay Us For:
Good advice, all around.
A friend of mine, who is an experienced law teacher, says that the way he gets through exam-grading season is by reminding himself that teaching and writing are so much fun, he'd be willing to do it for subsistence wages. So, everytime he grades an exam, he thinks to himself, "there! I just made [annual salary / number of exams] dollars!"
Posted by: Rick Garnett | Apr 27, 2006 2:32:37 PM
As a 1L, I agree that professors shouldn't make the exam too easy. I feel a sense of comfort when I read a question that brings scores of issues and ambiguity to mind, because I hope that my classmates are missing things. Alas, they always miss less than I think they will. But a real softball is frustrating, because there's no way for me to distinguish myself based on knowledge of the law alone.
Posted by: Fingerprint File | Apr 27, 2006 5:54:57 PM
Yesterday I took what I think was the fairest exam I've ever taken, law school or otherwise. It was fair because although the professor clearly had very strict demands (tight, tight word limit--1600 words--and rigid definitions of what constituted a good answer), said professor was very straight forward about their expectations. They provided sample examples with model answers and the exam tested in a manner similar to the presentation of material in class. It was also open book and open note, but this is incidently. The most important thing is to make clear what expectations are, and how points are awarded. I cannot see any reason why not to make available sample exams with model answers. This is the best way to level the playing field so that it is not a game of who knows someone who had the professor, or who will suck up to the professor and weasel out details about the exam and so on.
I'd go further. If I was a professor, I'd ask the students who got the three highest grades in the class to give me their notes and outlines for distribution to the next semester's class. Those students would then be on notice that there was no magic bullet in the form of a previous outline and that to get an above average grade, something more was required. Any volunteers for trying this?
Posted by: Bart Motes | Apr 28, 2006 9:23:32 PM
It's great that students are commenting on this thread, and I would love to hear more about what students thinks makes for a fair exam (or for fair grading). I believe my school is typical in that student evaluations are done before the exam, so profs. don't get feedback in any systematic way on their exams. And of course, when students are walking out of a 3-4 hour exam, tired and anxious, it's not likely that they will say to the professor, "Well-crafted test, nice job!" and even less likely that they will anything critical about the exam to the prof's face (except for a half-joking remark about how long/hard it was).
Second, as to Bart's suggestion, I have mixed feelings. On one hand, while I try to discourage students from relying on old outlines, I'm always amazed at how much students do anyway. That's one reason that I usually test on any significant new case that has been decided since the last time I taught the class, but that doesn't solve the problem entirely.
On the other hand, I'm not sure your suggestion would further the goal we both seek. First, I might not easily get permission from those top students. Second, despite my making the point that this WASN'T enough, I fear that a good number of students would be likely to say "cool, I got the outline from the A student! I don't have to do one of my own, or even study as hard, I'll just work off of this!" Maybe that instinct is wrong, but as much as I say "don't rely on old outlines, Gilberts, etc.," I still see students doing it on a regular basis.
Posted by: Joseph Slater | Apr 29, 2006 10:39:54 AM
As a law student, I can say it is completely rational for students to rely on old outlines. Obviously one shouldn't rely on them exclusively, but it's certainly helpful to bring together old cases and doctrines that the professor teaches in the same way from year to year. Most professors keep at least 80% of their courses the same from year to year, I can say from personal experience. I don't even really see why professors object to it; law classes, espeically those using the socratic method, leave things very unclear. Much easier to have the three-part test that we can actually throw out there on the exam.
In terms of fair exams, I agree that past model answers are extremely beneficial to students. Different professors want different things, why make us read your mind? Intricate issue-spotters with complex, head-scratching issues are the fairest exams, in my view. Also, hard exam with lots of time is fairer and more rational than easy exam with desperate typing.
Also, do not give overly restrictive word limits. I find it retarded that half the class spends half the exam cutting out words like "the" and changing "do not" to "don't" on some of these exams. It's just silly and not a good measure of our skills, especially when there are a ton of issues to cover.
Posted by: Guest | May 3, 2006 11:50:35 PM