Archive for October, 2014
Oct
31
Once upon a time, a long long time ago in a desert far, far away, I took a Physics class. The professor used one of them thar new-fangled personal computer thingies (pre-Apple, pre-IBM PC) to manage his grading tasks. He also introduced a solution to the “one genius makes the rest of us look bad” problem.
For the uninitiated, the “one genius” scenario is the case where a test is graded on a curve and just one person aces the test while everyone else bombs it.
His solution was simple: normalize to the second-highest score.
I have adopted this approach for any of my classes with ten or more students. If there are less than that, I’ll normalize to the highest score.
Here’s how it works. Let’s say a test is worth 200 points. A genius scores 190, but the second highest score is 150. What I will do is add 50 points (200 – 150) to everyone’s grade. Everyone, that is, except for the genius-she will get 10 points to bring her grade to a perfect score of 200.
The rationale is simple: If two or more people do well, then my teaching methods seem to be working, but if only one person does well, we assume that person is a “genius” and treat their grade as an outlier to be ignored.
The bottom line: it’s my job to help you succeed, so it’s my responsibility to grade fairly.
Note: Technically I’m applying an offset rather than normalizing, but you get the idea.