Okay, so I'm not drinking quite
yet, and I've got a question for all the more experienced teachers and graders out there: curving.
Do you do it? And if so, when and how?
Here's the deal. I've now graded all the midterms from my one survey section (the better one), and I'd initially been thinking about scoring them on a slight curve--say, calculating grades out of 95, the highest grade, rather than out of 100. This would give everyone a small boost and perhaps increase goodwill: I know I'm a pretty hard-ass paper grader, and I'm perfectly willing to be more leniant on the midterm.
However, in the end, I wound up with 5 or 6 grades in the A range (out of 30 total students), which seems reasonable except insofar as the midterm really wasn't that hard. If I curve, I'll have probably 9 midterms in the A range--but I'll also boost a few students who shot themselves in the foot in dramatic and tragic ways and whom I'd like to be able to give, say, a B- rather than a C+.
Justifications, rationalizations?
The other issue here is that I have another section still to grade, and while I think it possible that a couple of students in that class might have hit 95, there are likely to be fewer As and a lower overall range. I don't feel comfortable curving one class if I don't curve the other one, but I wonder if there's a rationale for curving only
within a given class--that is, curving the one class out of 95 and the other out of, say, 92, if that's the high grade there. Sometimes I think that makes transcendent sense, and other times I think it's absurd: a given student can't help the fact that she's in one class rather than the other.
So. I throw it open to the fray!
-
phd me commented at 8:03 PM~
Well, having accepted that fact that grading remains somewhat arbitrary no matter what I do, I don't curve. My justification, right or wrong: the exam/essay/project is a "fair" assessment of their abilities; the students knew the expectations from the beginning; I'm available for help at any point leading up to the assessment.
I'm not a complete ogre, though. I allow rewrites on papers on a case-by-case basis, and I incorporate multiple types of assessment to hit on multiple student strengths (no exam worth 80% at the end of the semester, thank you). So far, I haven't had any true complaints and no requests for extra credit, so curving hasn't been an issue for me - yet.
Good luck!
-
Rhonda commented at 10:06 AM~
I don't curve, either. Instead, I'm just a little generous in my rounding at the end of the term. Our university has a liberal drop policy, so my final average tends to skew in the B or B+ range.
I do allow rewrites on a case-by-case basis, and I used to grant bonus points for high attendance.
-
BrightStar (B*) commented at 3:22 PM~
have not curved, but have graded on improvement over time.
-
commented at 6:55 PM~
I don't think there's anything wrong with different curves for different sections, but I would *definitely* avoid curving relative to the extreme values like highest grade.
-
commented at 9:32 PM~
Belatedly - I don't curve either. I have done what tazjia says, dropped a question if everyone bombed and there was clearly a problem my teaching of it, but really rarely. But then, math is not my strong point, so I figure, why complicate things? ;-)
-
Professor B commented at 5:08 PM~
I seem to fall in the middle on this as well. I suspect that engineering exams are more straightforward to grade than english midterms, if only because it is often very clear that students have no idea what to do. (Though I suspect that some english midterms exhibit this tendency as well) ;-)
In any case, I have typically curved an exam only in the sense that I will drop a section or assign a baseline number of points for a section that everyone or nearly everyone tanks. I try and design my exams to come out with an average of about 70-75%, and typically hit that or a bit better. I am still to lenient with grading, I think.
As an aside, when I did give the students in my last undergraduate class and 'open ended' engineering problem (one that could have multiple correct answers depending upon the solution they chose to pursue), the students in general hated it and did poorly, even though it was no more difficult than the previous exams. Just having the words "and defend your solution method" in the problem was enough to send them all into a frenzy of panic. That's hard to break.
-
Psycgirl commented at 7:30 PM~
I just don't *get* curving, so I avoid it at all costs (like New Kid, math is not my strong point, and I definitely don't want them finding that out when I mess up a curve!)
-
Bardiac commented at 6:35 PM~
I like the idea of curving IF you do it consistently and strictly, by which I mean, you say ahead of time that you're going to curve, and that you use a real curve (not necessarily a "normal" curve), where the average grade (mean, median, you decide) is actually a C.
I can see that sort of grading most easily in largish science/math type classes, where the grading seems more objective. The exception would be where, say, you've got just an exceptional class where SO many people wrote outstanding exams that they deserve As even though that would seriously mess up the curve.
For essays, I tend to think in terms of a C being average competence. A C does what the assignment basically asks, but doesn't do it especially well. (Because if an exam/paper/quiz doesn't do what the assignment asks, it's surely not demonstrating competence, and it's not "average" to me.)
Sometimes students mess up gloriously. In that case, looking for improvement between midterms and finals seems like a fair and reasonable strategy. But changing the grading scale just to save them from themselves seems unfair or something. They need to get an honest assessment of their performance, no?
