It’s finals week! That means it’s finals designing week! Lift up your hearts and be glad!

I’ve put together a few exams in my day, but it’s always a thought-provoking experience. A good test should be a mix of the difficult and the basic. But how tricky is too tricky, and how easy is too easy? In addition, there’s the balancing act of divvying up the points into sections (multiple choice, T/F, short answer, essay), and making something that’s manageably gradable.

Some of my considerations – as always, your mileage may vary:

  • Essays: I like these because they emphasize ideas over list memorization, but my snail’s-pace grading style (and having 55 students in class) requires me to limit them. Usually I stick to a single 10-pointer, with about half of those points tied to a clear grading rubric (for my midterm, the essay was a Potter Box ethics case study, and identifying the four steps was a point each).
  • Multiple-choice: Four options seems about right. If you’re using three, you might as well just do true/false, and using more seems like an unreasonable time-sink. The exception is if you’re one of those who includes a clearly wrong, or, god help you, joke option (I like using “The Spanish Inquisition“). However, keep in mind that “clearly” wrong is subjective – there’s always someone.
  • Short answer: I like these, but they always get away from me in the scoring. What started off a simple “list and explain the three things” inevitably mutates into a cesspool of fractional points. “Well, he named the term but his description wasn’t right but it kinda was in places …” Bleh. I usually err on the side of giving points to the student in these cases, but I should probably toughen up.
  • True/false: These remind me of Mitch Hedberg: “Have you ever tried sugar … or PCP?” I tend to avoid these because I never liked them, but I was always intrigued by the variant that adds “If False, explain why.” Some day I’ll give it a shot.
  • Extra credit: There are two flavors of this for me, which emphasize either “extra” or “credit.” The former focuses on the fine details I try to avoid in the test proper; my thinking is that if you’re actually in touch with the material at that high a level, you deserve a few points.The latter focuses on giving credit where credit is due: If I’ve been talking about something since day one, but it’s not specifically on the study guide, then it’s something you SHOULD know – and if you DO, well then, have a point! (use sparingly)
  • Difficulty balance: I don’t like to make questions that are hard for the sake of being hard – I save my obscure references for extra credit questions (if you want something extra, you’ve got to DO something extra). “Difficult” for me tends to refer to questions that don’t have a clear answer, or that require some reflection to determine which answer is best.Hard questions should be so because they require thought, not just memory. A good gauge is when a test-taker stops and takes that long gaze up at the blackboard, seemingly willing the answers to appear upon its surface (granted, some students do this for every question, so focus on your high performers).
  • Overall harmony: I tend to write more questions than will actually be on the test, then prune the leafy ones and trim those that are good but don’t necessarily fit or are redundant. This has the added benefit of providing a well from which to draw for future exams (if you’re the kind of sneaky bastard who changes up your exams, which I am).

This particular 35-pointer is almost done. Right now I’m seeing a problem with that point count, as it seems too small to fit the amount of material I consider appropriate for a final exam, but that’s what I promised, so that’s the framework I’ll need to work within. Regardless of design, however, I know that come 10 a.m. tomorrow, I’ll be locked in my office talking to myself about how someone could have chosen B – “Seriously? B? It’s obviously not B. How could you think it was B? …”