The need to change McMaster’s assessment structures The need for more distributed testing to improve student learning

Photo by Grant Holt

By: Brent Urbanski

University testing is the Neapolitan ice cream of student evaluation. With its two midterm tests and single exam, this format has become known to students as an imperfect trifecta. As this format continues to be observed by students across classes, it becomes clear that university testing needs to change.

Within McMaster University, infrequent heavily-weighted tests have become the standard. Their proponents argue that they offer a pragmatic solution to the demands brought on by an increasing student population.

This seems reasonable considering approximately 26,780 undergraduate students attend McMaster.

Yet, despite a relatively large student body, as many as 62 per cent of undergraduate courses have less than 60 students registered. Of the remaining third, another nine per cent is accounted for by first-year, faculty-core courses, where non-standard assessment methods have already been adopted. Thus, only 29 per cent of McMaster courses have rationale on the grounds of large class sizes for the current evaluative structure.

While the limited time that our professors possess is valuable and should be allocated appropriately, students pay an underreported opportunity cost that is similar in consequence. As students, university is intended to be a time to quench curiosity and inspire potential. This opportunity is contingent on our instructors’ abilities to teach.  

With the recent push by the McMaster Students Union to eliminate evaluations weighted 50 per cent or greater, it seems that a new horizon is bound. However, while removing grade-defining exams will function to ease student anxiety and diversify grade distribution, it does little to correct an inherently flawed system of learning.

Cognitive psychologists have known for decades that proper learning involves the deep and repeated consideration of material. The more you practice retrieval of information, the better your long-term memory will become.

Distributed learning involves learning material over time while interleaved learning involves practicing several units of content in rotation. Science states that adoption of these two learning techniques leads to resilient memory when combined with levels of deep practice such as testing.

For a full-time student, with roughly 62 days of content from five courses compressed into three testing instances per course, the present assessment scheme hardly encourages distributed learning.

Given the current structure, it is not difficult to understand why students cram for major evaluations. It has been shown clearly that cramming behaviour produces improved short-term results when compared to the long-term strategies of distributed and interleaved learning.

While long-term strategies promote lasting memories, a majority of students are hesitant over using them. The current grading structure is unforgiving, and students frequently resort to suboptimal learning techniques given the cost of failure. But as the goal of classes is to foster long-term retention of material, the university should diminish the value of cramming and reward optimal strategies.

To craft an ideal course, one must first break the association between testing and evaluation. After decades of experimentation, testing has been established as the strongest way to induce resilient learning; the average person, however, views constant studying as preferable.

What we need is more testing in university courses. Not only does this greatly improve student performance on final examinations, but also a majority of students claim that they prefer frequent, low-stakes testing in comparison to infrequent, high-stakes testing.

And even better, testing does not only reference closed-book, sit-down evaluations. The evidence for open-book testing, textbook homework, and take-home assignments is overwhelming. Any content that provides students an opportunity to elaborate on their knowledge in a distributed manner will produce worthwhile results.

At this point, there is no question that frequent testing improves learning, as we have seen with the recent success of blended learning. The major challenge lies within the feasibility of adoption. Will instructors and teaching staff take on the extra effort to provide their students with a framework for success? Only future transcripts will tell.  


Share This Post On