I know it's been a while and life took it's toll; however, after a change of adviser (from one to two so I'm guessing this is an improvement?) and a new research program on the way, I think I can finally get back to blogging!
Let's be honest, I was procrastinating when I came across this article by Matthew Franke from Howard University earlier this year in the Teaching & Learning Inquiry journal ("Franke M. 2018 Final exam weighting as part of course design. Teaching & Learning Inquiry, 6(1): 91-103"). To say the least this reading was revealing so let's dive into it.
Simply put, having a good old final exam does not really reflect students learning as much as instructor's choices.
Indeed, by choosing the final's weight, an instructor decide for which students the final will be relevant. As you can see in the figure below, a student with 89% pre-final exam would need to ace the final in order to get from a B to an A at the overall course (assuming the course uses the 10-point scale.
What does that mean? Only a marginal part of the student population has even an interest in doing well on the final. This phenomenon is called the ceiling effect. Conversely, the floor effect is the fact that regardless of the final grade, some students cannot get lower that a "floor" score.
Importantly, the ceiling effect is the one of interest since as educators we should strive to offer all students the same opportunities when taking our final exam. As of now, if an instructor assign a final exam worth 10% of the overall class grade, then, the lower your pre-final grade, the more you can benefit for the final. Inversely, the higher your pre-final grade, the less you benefit from the final. Extending this logic leads to situation where pre-final grade rightfully distinguish between a C (77%) and a B (87%) students, but the final exam provides a greater chance for the C student to reach a B than the B student to reach an A ... is that what we want?
There lays the two questions at the heart of this paper: first, what do we want the final exam to reflect? Second, are we ok with inequalities between students during the final?
Finally, the author proposes an elegant solution (among other things) to this issue. Why not treating the final exam as an extra credit assignment - hence optional (see table below)? Note that the assignment before the final then effectively becomes a "final" with the same weighing issues as the final had before. A quick fix to this last point is to weight that new "final" between 10 to 30 % of the total score (not less or more, see discussion in the article).
While reading this article I was first shocked by the realization that student learning might not be at all reflected by the final exam. Then, I was pretty please with the proposed solution (final exam as an extra credit assignment). Now, after writing this blog entry, I wonder if all that is not up to the instructor. After all, in some situations I might be OK with providing a greater chances to reach the next letter grade to students with lower grades pre-final. What I am certain of though is that transparency and communication between instructor and students are essential regardless of the strategy employed: eluding student stress toward the final exam and instructor time answering questions.