For those of you who don't know, in university-level math courses, professors like to choose a few types of questions that represent roughly the material that has been covered in the course, and then from those they select the most difficult questions possible of those select types. Being math professors, one would think that they should understand that this is not a way to accurately get a representation of students' ability and progress in the course.

Say I can perfectly do 90% of the problem types. Chances are that, on average, I will get about 90% on my exams. I might get lower than that by virtue of silly mistakes, and I will only get higher than that if that 90% I can do is represented by more than 90% of the exam questions.

Now say I can do 100% of the problem types, but only up to 90% of their maximum difficulty. This is probably where most students stand -- they know the concepts, they understand the material and can set themselves up nicely enough, but they get stumped by the trick questions. And in university-level math courses, it seems that every question must be a trick question, or at least an extremely difficult one. Therefore, on the average test, in this case, my percent grade will probably end up being the same as the percentage of questions of the exam that are at 90% difficulty or under, which is often around 50%. Even if the tests aren't multiple choice, sometimes you can get nowhere on a solution and end up taking a zero on a given question. This has happened on problem types at which I've been 90% proficient. That missing 10% of ability ends up costing 100%.

I understand that they want everyone to be the best they can be (so they push harder than expected), but I shouldn't have to know how to integrate the most complicated thing in the world in order to solve a problem that shouldn't really be testing how well I can integrate complicated irrational logarithmic functions in the first place.

What I think would be better is to put more weight on weekly work. A lab period every week could be used to give an assignment or test on a smaller amount of material, ranging from easy to difficult. This way someone could still manage a 90% on the test while being 90% good at 100% of the problem types. Exams would still exist, but they would certainly not account for 100% of the term mark, which may end up being the case in my calculus course this term. I don't have the time to put in the effort to become 100% perfect at every possible trick professors could conjure to throw my way.

I could and would throw some solid statistical arguments out there to show that this method of test-making is fundamentally flawed, but I'm too mad and worn-out from studying and doing assignment after assignment assigned as though every professor thinks that I'm taking just one course this term. Maybe somebody else can help me out with these statistics. Vince? Just kidding.

Labels: Brainstorm

Posted by Farewell, Free Will on Wednesday, February 14, 2007 at 4:19 p.m. | Permalink