May 10th, 2013

Exams: Maximizing Their Learning Potential

By:

We give students exams for two reasons: First, we have a professional responsibility to verify their mastery of the material. Second, we give exams because they promote learning. Unfortunately, too often the first reason overshadows the second. We tend to take learning outcomes for granted. We assume the learning happens, almost automatically, provided the student studies. But what if we considered how, as designers of exam experiences, we might maximize their inherent potential? Would any of these possibilities make for more and better learning from the exams your students take?

Review sessions—Some faculty don’t have in-class review sessions because that means one less period for covering content. The question is whether students benefit more from being exposed to additional material or from having a chance to organize, summarize, distill, and integrate the content they must now learn for the exam. Should students do this summarizing and integrating on their own as they study? Perhaps. Will they learn to do it better if they their efforts are guided by an expert who understands how the content domain is organized? Probably.

Typically in the review session, the teacher goes over important or challenging content. Students are supposed to ask questions and they do, but generally they focus their questions on trying to ferret out what’s going to be on the exam. There are better alternatives. The teacher who already knows (and loves) the content doesn’t need to review it. Students need to review. The period should be structured so that students are doing the work, with the teacher providing guidance. They can be working individually or in groups, but they should be solving problems, answering old exam questions, writing possible test questions, or extrapolating key concepts from assigned readings. Groups could be given different topics, concepts, problem sets, etc., and tasked with preparing review materials/study guides for the rest of the class. They could bring these materials to the review session and present and/or distribute them.

In addition to revisiting the content and seeing more clearly how individual topics relate, review sessions can also be used to help students figure out what’s going to be on the test. That’s a question they shouldn’t need to be asking the teacher. The answer is a function of being about to determine what’s most important and how the content is going to be applied. And that’s a skill students need to develop.

Exams—Regular exams don’t promote deep learning because the questions don’t challenge students to think. Many students memorize well; they forget with the same efficiency.

Questions that challenge students to think are much harder to write, and for that reason you don’t find a lot of them in question banks that accompany textbooks. The problem is not with the multiple-choice format per se. SAT and ACT questions are multiple-choice and many of those are quite challenging. If exams are returned to students, then new questions must be generated for each new class. It is smarter to let students have access to their exams (when they’re returned and subsequently in the prof’s office), but not to let them keep their exams. That way, questions can be recycled and across the years a collection can be developed, revised, and reused.

Exam circumstances rarely change. Students work alone without access to resources and under surveillance so that they don’t cheat or they cheat less. This newsletter does, with some regularity, highlight different kinds of exam experiences—like having the students take the exam individually and then take the same exam with a group. Their grade may be some combination of their individual score and the group score. Or let students prepare a crib sheet (of a specified size) that they are allowed to use during the exam. Preparing a crib sheet forces students to make decisions about what’s going to be on the exam.

Debrief sessions—Typically teachers go over the most missed questions, but that approach may not be the best way to maximize the learning potential that is still present after the exam. Teachers don’t need to correct the answers—students do. Whether in groups or individually students can be given the chance to find the correct answers and fix their mistakes. Maybe that happens during the debrief session, or maybe students do the work at home, completing it before the next class session. Maybe their grade isn’t recorded until they’ve corrected their errors, and maybe it’s a few points higher if they get all their mistakes taken care of.

Debrief sessions can also be designed so that they address some of the decisions students have made about preparing for the exam. Class attendance makes a difference. You can say that, but you should show some evidence. Take the five highest exam scores and list the number of times that group of students missed class. Take the five lowest scores and list the number of class sessions that group missed. Let the facts speak for themselves. Many students aren’t taking enough notes in class. You can say that, or you can demonstrate it. Pick a question that many people missed. Identify the date that material was covered and have everybody look at their notes. Do they have what they need there to answer the question? Were they absent and got notes from somebody else? Do they understand those notes? Quick discussions of topics like these can be concluded with students writing themselves a memo addressing “things I learned taking this exam that I want to remember for the next one.” Collect those memos and return them shortly before the next exam.

Exams motivate students and learning results. They review their notes, read the text, and talk with each other. The question is how much and how well do they learn? How seriously they study determines part of the answer to that question. But it is also answered by the design of the exam experience including what happens before, during, and after the event. Exam experiences can be designed so that more of their potential to promote learning is realized.

Reprinted from The Teaching Professor, 26.4 (2012): 3.


  • Pingback: The LSU Top 5 #31 | lsuvietnam()

  • Jim Benner

    I am looking for material on test-making with Student Learning Outcomes and higher-order thinking in mind. In our course/program outcomes and in our "faculty-lounge" chats, we articulate the desire to see students do Application, Analysis, Evaluation, and Creation (i.e. higher order thinking on Bloom's taxonomy). Then, we build too many tests that never get past Remember or Understand. This is a dual challenge: creating tests that match our learning outcomes and ask students to think at higher levels. Of course, other assessment methods can do some of this, but why not tests?