Having the Final Say
Rob Weir offers tips on designing final exams.
December is the cruelest of all academic months. At exactly the time one is running out of steam, a torrent of work descends, and holidays rain down like locust plagues. There’s Ashura, Hanukkah, Festivus, Christmas, Kwanzaa and First Night. Heck, even Beethoven’s birthday is in December. There’s so much going on that I'm personally thankful that Hindus held Diwali in October this year. Forget sugar plums; getting through December induces visions of sleeping for a month.
So how does a new hire get through finals with sanity intact? The temptation is great to design a multiple choice final, have your students fill in bubble sheets, feed those babies into a Scantron machine, record the grades, and hide in your closet until spring semester begins. It's not the most pedagogically sound method, though. Let’s start with what not to do.
Don’t exact revenge. Maybe you had a class (or two) in which students were inattentive or disrespectful and you’re thinking of lowering the reality boom. Sorry; you should have done that earlier in the semester. Maintain professional poise and design a final appropriate for the subject matter, not your revenge fantasies. When I was an undergrad I had a new and very bad sociology professor who didn't like student questions, even when she was flat out wrong. She gave an essay final, calculated to the third decimal point how low she could mark the papers of those who challenged her and drop their grade by a letter, and proceeded to do exactly that. Result? About 75 percent of her class successfully appealed their grades and she was gone.
Don’t experiment with your final. Experimenting with teaching methods and evaluation methods is a good thing, but the final isn’t the place to dust off an intriguing new idea. You should hew closely to whatever methods you've been using to evaluate throughout the semester. If you've assigned essays all semester, don’t suddenly opt for an objective exam; if you’ve required engineering students to work with theoretical models for 14 weeks, don’t ask them to design a bridge in week 15 unless you’ve told them weeks ago that this would be the final project and the theory they've been working on is germane to that goal. In an earlier essay I spoke of the importance of viewing one's syllabus as a contract with students. You laid down the conditions, so honor them.
Don’t use the final to test things you never got around to teaching. The first time I taught the U.S. history survey I was so raw that I barely made it to the 1960s. But I wasn't so green that I failed to recognize that most of my community college students needed guidance to comprehend nuances, interpret primary source documents, and apply conceptual models. I simply couldn't tell them that they were "responsible" for the remaining material, so I designed a final that covered a smaller piece of historical turf. The moral is that the professor controls the agenda and it's unfair to expect students to paint on walls that never got built.
Having said all of this, one should also resist the temptation to follow the path of least resistance. If a final is worth giving at all, do it right. A few tips:
Make the final harder (and probably longer) than other exams. This is the culminating exercise of a course and it's fair game to ask students to demonstrate mastery of the material. It should have a cumulative aspect to it, be it in the form of actual detail or in (forewarned) application of central concepts. I am currently teaching a writing seminar that's themed around critical thinking. Students are expected to demonstrate mastery of a list of distributed elements of critical thinking (that were previously discussed and practiced). A computer science friend who teaches programming requires each student to write an original program. They must burn it onto a disc and submit it. If the program does not run, they get an F (but one more try at making it work). If it does run, they get an automatic C and mark ups are akin to scoring Olympic skating — one gains points via degrees of difficulty and elegance of execution. A final should separate the wheat from the tares. Made sufficiently (and appropriately) hard, it will help you assemble your final grade distribution. (If it doesn’t, make it harder next semester!)
Multiple measurements will yield better results. The computer science professor I mentioned doesn’t score his final on the program alone; he also asks students to write a short essay and has several other exercises as well. Varying what you evaluate is good practice for all exams, but the longer time slots most colleges allot for finals provides opportunities to spread the points out even further. The humanities have long favored a model in which students write a major essay, a shorter one, and perhaps do some identification problems. The sciences often divide finals into written and lab demonstration exams.
Prepare your students. Give students advance warning of how the exam will be set up. If you can, provide practice exercises. You might, for example, use the course Web site to post a model of a well-done essay or a well-crafted theory application. You could post a video of a successful experiment. In 100-level classes, it's not a bad idea to create a finals study guide. When I've taught at colleges whose students come to college ill-prepared, I've gone a step further. I wrote three essay questions and advised that two of them would appear on the exam exactly as written, but they would not know which two. I repeated this with identification questions, sequencing problems, and other materials. Spoon feeding? Perhaps; but students had to study for twice as many items as would actually appear and I bargained that they’d learn something in the process.
Make instructions and expectations crystal clear. However you configure the exam, make certain that students know what they'll be doing and how you'll evaluate it. Don't, for example, let students think they can blow off the essay portion if the objective section is only 25 percent of the grade. If you have any special conditions, spell them out in person and in writing. For example, when I give blue book exams I forbid the following: cell phones, laptops, notebooks, and backpacks. I do, however, sometimes permit students to bring printouts of primary source documents (which are checked upon entrance).
Make students aware of your college’s honors code. Yes, it’s in the handbook and, yes, you’ve mentioned it scores of times. Remind students again. Many professors — especially those whose final assessments are unmonitored — print the honors code right on the exam. Good idea!
If you can, break away from conventional exams. I'm aware that some colleges have non-optional requirements about finals. If you are allowed leeway, however, consider alternative examination methods. I do not believe for a second that blue book exams are necessarily the best way to measure student success, and I’m a near-total skeptic on the "measured outcomes" fad sweeping academe. (Way too much of the latter is jargon-ridden gibberish fashioned by administrators who don’t teach to mollify demagogic politicians who don’t think!) Okay, take your potshots at my iconoclasm, then ask yourself if there is a one-size-fits-all way to assess what you teach. Of course not! If you’re teaching an upper-division course, the best assessment tool may be a research paper, a lab demonstration, an oral presentation of research, a piece of creative fiction, a collaborative project, a demonstration of craft, or a work of art. For some advanced history courses I require students to present their research orally, submit a formal paper, and design a Webpage. But even in some 100-level classes I’ll opt for a take-home exam over blue books. If you do the latter, redouble your warnings about the honor code and make darn sure you check the work and punish transgressors. Also make sure that students know exactly when the project is due and entertain no excuses to the contrary.
Take your time grading. The best way to stay sane in December is to not drive yourself over the precipice upon which you’re already standing. There’s nothing wrong with taking as much time as you're allotted for submitting grades. Alas, I’ve taught in a few colleges whose idea of "enough time" is four days after finals end. That's insane and a battle worth fighting with administration, but calculate the time math and pace yourself accordingly. If you’ve got a hundred students and two weeks to grade, do ten exams a day and the weekends are yours. That leaves plenty of time to contemplate Festivus grievances!
For further consideration:
1. A law student’s study tips provides inferential ideas on how to write a good exam.
2. Psychologist John Ory shares his thoughts on finals.
3. Follow the PrawfsBlawg discussion on take-home finals.
4. See a University of South Florida computer design final exam.
5. If you can access Questia, see Ian Brown’s observations in "To Learn Is to Teach is to Create the Final Exam." (It’s also in College Teaching, Vol. 39, 1991.)
6. A Berkeley professor's prep guide for students.
Search for Jobs