The exam results were mixed. On one hand, the average (70.8%) was not that bad, nor was the standard deviation (10.6), but the distribution shows a strong tail to the left:
I will probably end up scaling the exam upward by about 5 points, I will decide after more detailed analysis and let you know later this week. No one got 100%; 18/20 was the maximum score.
There is one prevailing theory to explain the average being a bit lower than expected, given that I thought the exam was far easier than the homework sets. Below is a scatter plot of homework scores (x axis) versus overall grade (y axis). There is a mind-bogglingly strong correlation between overall grades and aggregate homework scores.
So. If you do the homework, statistically there is a strong tendency for you do well in the class. Part of this is due to my tendency to reuse questions (in spirit, if not precisely). Part of this, however, is the fact that the homework assignments are constructed to get you thinking (and reading!) about the most important topics. It is not busywork ... I promise that if you grind through the homework, it will pay off. If you find yourself having trouble with the homework, you can always email me or drop by my office.
Anyway ... two main points: (1) the test will be scaled a bit, and (2) the homework really is worth your time, as much as it sucks.
No comments:
Post a Comment