Instantiating Specifications Grading in Computer Science Courses
David L. Largent · Proceedings of the 53rd ACM Technical Symposium on Computer Science Education V. 2 · 2022
Historically, I assigned points to assignments, and determined course grades based on the percentage of points a learner earned. This meant that a learner could do poorly on some assignments and great on others, and effectively generated an "average grade." One could not know, given a grade, what an individual student could or could not do. Also, the method of counting points inevitably leads to students begging for partial or extra credit. Seeking a grading methodology to authentically reflect what a leaner accomplished, I explored the use of specifications grading (SG), as described by Linda Nilson (Nilson, 2015). Basic tenets of SG include providing a very clear, detailed description of what is expected (the specification), and then evaluating submitted work as complete or incomplete against that specification-they either met the specification, or they did not. Simply providing clear expectations can make a significant difference in a learner's ability to submit what their instructor expects. I evaluated assignments as complete/incomplete in a variety of courses. The course grade was based on how many of each assessment category the learner completed. For example, if a learner passed at least five of six quizzes, they earned an "A", provided they had an "A" in all other categories. Safety net provisions, like having a very limited number of "oops bits" learners could use to resubmit an incomplete assignment, were provided. My experience has been mostly positive. I will present changes in how I evaluated assignments, and my positive and negative experiences doing so.