The U of MN exams I give have 5 multiple choice questions on them. Usually I give a scantron so I can easily do an item analysis, but I forgot to hand them out with the exam this time around, so when I handed the test back I asked students to tally their responses. The visual was a pretty neat way to analyze the results!
As I was grading this years Exam 1 in my concurrent enrollment college level physics class I had a feeling that students were really rocking it compared to last year, so naturally I spent 1-2 hours on a Saturday night playing with the data. This box plot is, I think, the best representation, showing that all four quartiles increased. This is verified in that the mean had a statistically significant increase (p<0.05) and the standard deviation decreased from around 13 to 9.
I am in the process of writing a blog post about why I believe this increase of scores occurred, so you’ll have to wait a bit for that. The teaser is that it’s a kind of Standards Based Grading (SBG) implementation for a class in which I can’t really do SBG.
UPDATE: Here’s the blog post, When You Can’t Do SBG