Grades and Evaluations

More than a decade ago, I found myself newly minted as a department chair. When it came time to do annual evaluations of faculty (all of whom were more senior than I), I sweated bullets over it. I tried to systematize it as much as possible in the name of objectivity. Whether I was successful or not is questionable, but I survived without a mutiny at least. One of the indicators I looked at was grades assigned versus evaluations received. I made a scatterplot of the two variables and looked for outliers. This had to be done by hand, laboriously typing in numbers from printed pages.

Flash forward to the present, where I built an electronic system to store evaluations. Mass comparisons are now only a few queries away. I finally got around to doing this a couple of weeks ago, and the result is shown below.

The student ratings for "The instructor appears knowledgeable and competent in the field" appear on the left, with 1 = strongly agree and 5 = strongly disagree. The bottom axis is grade point average on a four point scale. Each dot represents one faculty member (with at least 100 ratings). Although there is a bit of tendency for higher grades to go with better ratings, the correlation is pretty low at -.22. There is an obvious ceiling effect at the bottom of the graph. Many students are giving top ratings across the board.

I expected the correlation to be considerably more significant. The good news is that I was wrong--grades don't seem to be being given away in order to get good evaluations.