A question on exams
My friends, please note the time of this post. I have just finished posting my final exams on Blackboard. Finals aren't till next week. They are three-part exams. A take home document analysis, in-class IDs and in-class essay. I was not given document questions on exams when I was an undergraduate. We had basic questions (compare/contrast, causes/effects of X, trace the development of ...)and were expected to refer to documents we had read. I never really had to do document analysis in any depth, even in papers, till I got to grad school. Even then, not so much, except in the context of research papers. So why ask the students to do it? Because it's interesting and they should know how to read and analyze primary sources? Because it's good for them to know how professionals know what they know? I think those are good reasons. And because we work on these things in class, it makes sense to assess what they have learned.
But the other part ... I guess there were profs who gave us lists of IDs to study. Maybe. And sometimes general topic lists. But no one ever gave me the questions in advance. So I give them four questions. Then on the day of the exam, I choose two. The students write on one. But I have to consider this -- the essays aren't all that much better than when I've given the questions blind and on the day. In fact, only a couple are hugely better, and the rest not so much. This was true when I gave both essays as take-homes. The essays were a bit better organized, but not noticeably better written or proofread. So why am I staying up so late, a week before exams, to give them the exam questions in advance?
Does anyone else do this?
And, on another related note -- every time I look at exam questions I might think of pinching from old exams, I think ... hey, maybe I should remember these, and my course would be more focused. Does anyone else do this? If you do, how do you walk the line between crafting a course with several clear content goals and teaching to an exam? BTW, I'm not totally clueless, just very tired. There just aren't that many questions we ask in surveys, anyway -- just multiple creative way of asking them. So it should be easy not to "teach to the test" while still having a clear idea of course content outcomes ...
