One complaint you hear fairly often from students about professors or courses involves exams whose relationship to material covered is vague or unclear. Sometimes students feel like they cannot reasonably guess what terms they will be asked to identify or define, that such terms may include ideas never discussed or reviewed in the class or in extreme cases, only minimally discussed in textbooks.
The final exams in my survey classes are one of the few places that I ask students to demonstrate command of a concrete body of discrete facts. I divide the exam between identification questions, which are 40% of the grade, and an essay, which is 60%. Usually I have eight, sometimes ten, items that I want students to identify.
To help them prepare, I hand out a sheet with around 40-50 terms from which I intend to draw the terms which will appear on the exam. Quite a few of them overlap, and I suggest to the students that it would be a good idea to think about incorporating the overlapping or related terms into a single identification if one term from that cluster appears on the exam. E.g., if you’re asked to identify “Igbo” and “Arochukwu” and “title societies” also appear on the study list, use your knowledge of all those words in a single answer.
This seems to me to be a really good strategy for ensuring that the students leave the class knowing a set of concepts. They know what to study, and I draw my identification questions from my lectures and from the reading. Even if they skipped a reading earlier in the semester, this motivates them to go back and do it seriously now.
I was surprised a few years back when talking with a colleague about this approach that this colleague saw this as giving the students an “unfair advantage”. My colleague felt that if identificaton questions weren’t potentially drawn from any and all course materials without any hints about which materials, the students would just cherrypick the few highlights I was directing their attention to, and be rewarded for not fully engaging the course material evenly over the whole of the semester. This seemed like a really weird argument to me. With finals approaching here, I’m curious about what the larger academic universe thinks. How do you construct finals in courses that are intended in part to teach discrete bodies of facts and information?