In the ed biz, it's time to grade like a demon and yes, write final exams. (Unless you're lucky enough to be completely finished with your term, in which case, you're not reading this anyways.)
Depending on the course, I have a couple different exam formats I write. Whichever format I use, I have a couple goals for every exam, some of which I achieve, and some I just aspire to.
From most mundane to most important:
I try to write an exam that a student who's never bothered coming to class or doing the reading will fail miserably. (Yes, sure, someone who's already expert in the field should do just fine, but my students don't come into my classes as experts, or they wouldn't be in my classes.) (I think I generally succeed pretty well at this part.)
I try to write an exam that a student who's come to class, done the reading, and put forth a solid effort in preparation and studying should be able to show what they've learned and pass. (I generally do okay here, too. An exam isn't about showing that I can trick people, or that I'm smarter than they are. I could trick them; what's the point? And I'm not smarter than many of my students, so trying to demonstrate that I am would be a waste of effort.) As a corollary, I want a really stellar student to be able to show how stellar s/he is.
I try to write an exam that will be a useful learning experience in and of itself. I want it to help students put together the information they've been learning so that it makes better sense to them (a goal of the comprehensive essay part). In an ideal world, the student who's been working hard all semester will realize through preparing and taking the exam that s/he's learned a whole lot, and that the hard work was totally worth the effort. (I don't think I succeed here too often.)
The exam style I use most often for Shakespeare or lit classes looks basically like this:
Short identification section - terms, concepts, dates, etc. Write a 1-3 sentence definition and connect the term/date to something we've read or discussed in class. There's always some choice in this section, do 6 of 7 or something, so that someone can forget a term and still do well.
Passages - from whatever text(s) we've been working on. Identify the text, speaker, and context in a few sentences. Write a short explication of the passage focusing on some aspect of word choice, imagery, concepts, poetics, etc. Don't paraphrase, but do feel free to make connections to other parts of the text, or to other texts from class. Again, there's always some choice in this section.
The final looks the same, except there's also an essay section which tries to get students to make connections across the whole semester. Again, I give two essay choices, and each student chooses the one s/he wants to write.
The format, of course, isn't something I made up myself. Nope, in fact, I adopted it from my graduate school mentor, for whom I TA'd a number of classes (as well as grading other classes). My graduate school mentor adopted it from his graduate school mentor. And I'm guessing that person adopted it from her graduate school mentor, though I can't be sure.
So, writing my Shakespeare exams, I have an academic grandmother who I've never met (and probably never will), but I use her exam format.
Ideally, I suppose, adopting teaching techniques and strategies comes from the apprenticeship aspect of graduate school teaching. (At any rate, I don't remember any pedagogy classes teaching me to write exams.) The exam format works well enough for me in general that I continue to use it. It's somewhat a pain to grade, though since my grad school mentor and his were both in PhD programs with lots of grad students available for grunt work, I suppose they didn't much worry about that. And it's no worse to grade than a lot of other formats I've seen.
I was talking to a colleague in another field recently about exams; she's in a field where multiple guess exams are pretty standard, and so she gets to grade at least part of her finals via a scantron machine. I realized that I'd have no clue how to write a decent multiple guess exam. But she learned through her apprenticeship not only how to write them, but that they're appropriate to her field.
Since my apprenticeship was moderately long (3+ years of TAing in various ways, and more grading for additional classes), I had lots of opportunities to pick up exam formats, watch others teach (which, of course, I'd been doing since I was 5 years old). And there are other formats I use occasionally, especially when I want to challenge students in a different way. One of the best is to give students a group of passages from texts, and ask them to explicate one fully, while referencing another. That one works well as a take home exam, especially in a class where there's a fair overlap of conceptual or stylistic work, such as a single author type class.
Even though I picked up exam formats from my various exemplars in graduate school, we almost never talked about exam writing as such. I didn't think to ask most of my professors why they wrote exams the way they did, or what they wanted to accomplish with exams. I just wasn't "there" yet, wasn't ready to think about teaching that fully. It's a shame, really.
And my graduate program really didn't encourage students to talk about teaching as such, to talk about what we wanted our students to learn or reveal by taking exams. (We did, on occasion, talk about such things, but not out in the open.) Our professors rarely talked to us about teaching per se, either.
Now, when I'm so much more ready to think about strategies for teaching, I'd love to go back and ask the best teachers from my past about their strategies and their exams. I wonder, though, how conscious some of them were about what they were doing? Or how much they just adopted their mentors' formats, assuming that the mentor had thought things through better than they could?