Learning outcomes assessment ratings scales and their interaction with rubrics in Canvas

The default scale in Instructure Canvas for learning outcomes uses Exceeds expectations (5 points), Meets expectations (4 points), Does not meet expectations (0 points).

This yields rubrics that look like the following.

If a student meets all expectations, but does not exceed, then the student would obtain 6 out of 10 points. Used as a grade, this is a D- at 60%, hardly equivalent to meeting expectations. A borderline F. Note that mastery was set at 3, so the student has mastered the outcomes and has a D- for a grade based on the rubric. 

A shift in the scale and terminology provides a more flexible and realistic alternative.

Here the word Sufficient refers to a "sufficient demonstration of mastery of the learning outcome." Mastery is set at 4 points, Sufficient, and produces a rubric that appears as follows.

On this rubric Sufficient mastery produces a more realistic 8 out of ten possible points for an 80% score. The word Optimal refers to the student having made optimal choices in solutions, optimal demonstration of the outcome. Suboptimal might be a partially demonstrated mastery or a developing level of mastery. Suboptimal control of grammar and vocabulary. A suboptimal solution to a problem. The student built a cabinet and in general the cabinet is well made, study, but the doors are not quite hung properly, that is a suboptimal solution to "The student can build a cabinet." A fully working cabinet is sufficient. A cabinet that includes useful additional flourishes like a spice rack underneath the cabinet and hooks for coffee cups might be a more Optimal solution. 

No discussion of learning outcomes would be complete without the issue of whether a learning outcome is binary: the student either can or cannot demonstrate the specific knowledge, skill, or value. This only works at the lowest level of outcome. At aggregated levels such as the course level gray areas appear. Outcomes such as Stats 3.0 above are no longer binary all or nothing demonstrations because they are made up of many lower level, perhaps dozens, of much more specific outcomes. 

One other note on the ratings choices rejected. At some level I expect all of my students to excel. My expectations are that the students can and will exceed my expectations. In other words, I have personal or possibly semantic challenges around evaluating based on expectations. Expectations is setting a future location that the instructor expected the students to get to, and if the student does not, then the student did not meet those expectations. The student's performance becomes dependent on the expectations held by the instructor - not on something the student can control.

Optimal and Sufficient rate what the student has done without any preconceived expectations. Optimal, Sufficient, and Suboptimal look back at the student has done, not forward at what the student was once expected to have done. "The student chose an optimal community service project and made optimal choices in set up, and delivery. The final report was sufficient." Optimal, Sufficient, and Suboptimal provide more control, more agency, to the student. 

Post-script 24 July 2021

A table was developed as a mapping guide among rating scale systems.