The Grade Book That Lies to My Students

I do not use weighted categories in my Canvas gradebooks. I never have. The math behind simple point totals is fully transparent: a student earns 850 out of 1000 points, and that is an 85. No hidden formulas, no intermediate normalization steps, no categorical arithmetic that requires three operations before you arrive at a meaningful number. A student who wants to dispute a grade can sit down with a calculator and check the total points in under a minute. That is the system I want.

Canvas does not want that system. Or rather, Canvas does not care what I want. Its gradebook is built around assignment groups as the primary organizational unit, and its student-facing interface surfaces percentage summaries regardless of whether you have enabled weighted grading. So even when my course runs on pure points, my students see columns organized by group, percentage breakdowns by category, and a visual presentation that implies weights exist. The underlying math may be correct, but the interface tells a different story. I cannot turn this off.

This bothers me more than it probably should, and I have thought about why. The issue is not cosmetic. When a student sees a "Participation" group sitting at 72% next to an "Exams" group sitting at 88%, they naturally ask how much each one counts. In a weighted course, that is a meaningful question with a clean answer. In my course, it is a question the interface invites but cannot honestly answer, because the groups are not weighted at all. The percentage figures are real, but the framing around them is misleading. A platform designed to communicate grade structure is, in my course, actively miscommunicating it.

There is a genuine argument for weighting that I do not want to dismiss. Weighted categories give course designers a stable architecture: if you add two more homework assignments mid-semester, the homework category still counts for exactly 30% of the final grade, and nothing else shifts. With a points system, every new assignment changes the denominator, which means it changes the implicit weight of everything that came before. If your course structure is stable, this does not matter. Mine tends to be stable. But the weighting approach is more robust to the kind of mid-course adjustment that happens in real teaching. I understand the appeal.

What I do not understand is why that robustness requires opacity in the student-facing view. The two goals, design integrity and computational transparency, are not actually in conflict. A course could show students the running point total, the total points possible, and the implied percentage, all without requiring them to reconstruct categorical normalization by hand. Canvas could offer this. It does not, at least not as a clean default. Instead it presents a percentage-based view that fits weighted courses naturally and fits mine awkwardly. The workaround I use, a single assignment group with explicit point totals and a plain-language explanation in the syllabus, is functional but inelegant. It is the kind of fix that should not be necessary. Instructure Inc., are you listening?