November 14, 2017

Mastery grading: How one professor uses tests to help his students learn

Mark Blair, an associate professor in the Department of Psychology, rethought the role of tests in his course. The result has been a dramatic improvement in student engagement and learning.

Mark Blair (associate professor, Psychology) has been teaching COGS 100, a cognitive sciences Breadth course, to roughly 100 students per semester for the past 12 years. A review of class results over the past few years convinced him that students weren’t learning as effectively as they could. Grades were static or even declining, and the practice quizzes he gave to students to prepare them for tests didn’t seem to make much difference. In fact, when he repeated quiz questions on his tests, students often responded with the same wrong answers both times. His observations led him to reconsider the way he handles assessment, and this fall he is trying a different approach called “mastery grading.”

Learn, test, learn some more

Mastery grading integrates student assessment into the learning process. Instead of writing a test once, receiving a grade, and then moving on, students use tests as a measure of what they know and don’t know, and retake each test until they achieve “mastery” of the material, defined in Blair’s class as a score of at least 90 percent. (See How Testing Is Done in COGS 100 below).

“It doesn’t matter that they fail. It only matters that they eventually succeed.”

“There’s no penalty at all, literally no penalty, if they fail [initially],” says Blair. “It doesn’t matter that they fail. It only matters that they eventually succeed.”

For Blair, the shift toward using assessments as a tool for improvement rather than simply as a measure of performance “at a particular time and place” is fundamentally important, and he has already seen a difference in the classroom: “That changes the dynamic. I’ve had so much more engagement from students … than I typically would.”

What really excites him is that his students now have the motivation and the opportunity to address problem areas.

“If you think about [traditional] assessment in university, there’s very little opportunity to correct your mistakes,” says Blair. “When you make mistakes, you don’t go back and fix them, you don’t figure out what it is you got wrong or didn’t understand, you just move on.

Now, he says, his students do go back, and he as an instructor can spend his time “coaching” them to better results.

Beyond the bell curve

Student response has been “really positive,” and Blair is enthusiastic about the learning he has observed: “I can tell you for certain that in 12 years of teaching this course I have never had a group attain that level of understanding.”

He points to a graph of test scores (see the examples below) that is strikingly different from the typical bell curve: “They always say there’s … 20 percent of students who just can’t learn no matter what you do. But I don’t think that looks like 20 percent to me. What that looks like is three.

“What that tells me is that those people at the low end, if you set up the structure right, they can thrive, they can learn!”

He’s not sure what administrators will say when he submits grades for a class averaging around 90%, but he does have a ready response for anyone who questions the results: “The proof is in the knowledge that these students are demonstrating and the work that they are willing to do.”

Related links

How testing is done in COGS 100

Mark Blair’s COGS 100 course covers eight major topics. For each topic, students write a test worth 10 percent of the final grade (the remaining 20 percent comes from a final test and various assignments).

The tests are available in the Canvas course shell, but access is restricted through an access code to a 20-minute block of time built into almost every class. During that time, students can take or retake tests until they achieve a grade of 90 percent on each one. Between test periods they focus on the problem areas; Blair uses a flipped-classroom approach with videotaped lectures, which gives students the opportunity to replay lectures on their own or ask for clarifications in class.

All test questions are multiple-choice and are randomly drawn from a question bank to minimize repetition during test retakes. The testing technology was implemented with the help of learning technology specialist Robyn Schell from the Teaching and Learning Centre.

Blair is quick to emphasize that his tests require critical thinking and don’t simply measure increased familiarity with the course or test material: “My test questions are designed on purpose [so] that that strategy just does not work.”

Student scores: The top graph illustrates how grades for the first test of the semester have converged toward the 90%+ range over time. The bottom graph illustrates the current results for the most recent test given. The wider distribution of grades indicates that many students are still working to master the material.