Specifications Grading: Restoring Rigor, Motivating Students, and Saving Faculty Time

I don’t like grades.

As a student, I oscillated between taking anything but superlative grades as a sign of my own failure and being utterly indifferent to grades as a secondary consideration to learning the material. Either way, grades were an imperfect motivator.

As a teacher, I am even more ambivalent about grades, which I see as something I am required to do in order to rank my students. I am always prouder of a student who struggles and reaches a breakthrough than the genius who coasts through the course, even though the latter receives the higher grade. My own experience as a student informs how I structure my courses, leading to policies that encourage regular engagement, choice in how to complete assignments, emphasis on the process over product, and often opportunities for revision. Each of these course policies marked an improvement, but they all retained the thing that I was in many ways least satisfied with: grades.

A few weeks ago a faculty development seminar introduced me to the broad strokes of Specifications Grading and since it seemed like the direction I have been moving my courses, I spent nearly an hour after the event jotting down preliminary notes for what that might look like in my course. At the end of that day I was intrigued, but needed more information. Over my spring break, therefore, I read Linda Nilson’s Specifications Grading: Restoring Rigor, Motivating Students, and Saving Faculty Time (Stylus 2014).

Broadly speaking, Specifications (Specs) Grading is a variation on a pass-fail, contract grading, and competency-based outcomes that ties course assignments to specific course objectives. This model, Nilson argues, has three major benefits. First, setting a high bar for “acceptable” work but giving opportunities for revision imposes rigor without making the professor into a jerk. Second, demystifying the grading process and offering flexibility reduces stress on the students. Third, eliminating partial credit saves time. Some model systems presented a fourth potential benefit of allowing teachers to give more of their limited attention to those students aiming for the higher grades.

In addition to an argument for its benefits, Specifications Grading serves as a guide to adapt traditional grading models to a specs system across two broad categories: outcomes and assignments/rubrics.

If you’re anything like me, you course outcomes won’t work for specs grading. Nobody ever really taught me how to write objectives so what I have in my syllabuses focus on what the students will receive. The conceit of an objective might be well-intentioned, but if the students can’t demonstrate what they are learning through the assessments, then it won’t work. Often this just means a subtle, but significant shift:

  • Students will gain a broad understanding of US history since 1877.
  • Students will be able to identify the major events of American history since 1877

Each of these objectives would then be demonstrated specifically by one or more course assessments. In Nilson’s model, some of these course objectives would correspond to basic, minimal standards like the one listed above. Students who achieve proficiency at those lower-level objectives would be able to pass the course with a C, while students at aiming for a B or A would have to also demonstrate proficiency at objectives that involve more complex skills.

The second part step involves developing detailed one level rubrics that explain everything that the assignment must have to be accounted “proficient.” Now there will be some variability in what that standard should be, but Nilson recommends building the rubric from everything you would expect to see in a roughly B+ assignment. When it comes time to grade the assignments, then, the assessment becomes a binary yes/no, along with some comments that might be used if, as Nilson recommends, the students get the chance for revision.

I have traditionally had an antagonistic relationship with most rubrics because most of the rubrics I have been required to use were a particularly poor match for how I wanted to grade such that someone who received 9/12 on the rubric was solidly in the B+ range according to how I grade. However, I found myself coming around to this model of rubric because it removes the splitting hairs and partial credits in favor of either showing that the students achieved proficiency or did not. The grade translation, in turn, does not come from an individual rubric but from how many assignments in which the student achieved proficiency.

and have been jotting down notes on how I can transform my existing courses with minimal disruption to anything but how I grade.

For my general education classes the assignments might look like (based on a syllabus for this semester):

To receive a “C” in this course (linked to the lowest tier of objectives)

  • Participation [in various forms] of 75%
  • Objective quiz score of 75% [I allow retakes and drop a quiz score, so I have exactly 2 students who are not clearing this bar right now]
  • Journals 10/15
  • Papers 5/5 completed, but not to “proficiency” with historical essay writing

To receive an “A”:

  • Participation of 95%
  • Objective quiz score of 90%
  • Journals 13/15
  • Papers 5/5 to proficiency
  • Completing a final project

The “B” range would obviously fall somewhere in between these two levels, with a “D” a little below “C.” The numbers might be off a little bit, but I would calibrate them based on what my final grade sheet looks like.

For my upper-level classes that are writing intensive and where the students complete three longer essays, a “C” may require revising one of the three essays to proficiency, “B” requires two, and “A” all three. For all of these classes, I am also toying with the idea of creating a list of “recommended” books for the course and allowing any student the opportunity to choose and review one of these books in place of one “proficient” paper—with guidelines for what constitutes an acceptable review, of course.

Specifications Grading also introduced me to a different paradigm to the student-teacher relationship. Students are not customers, Nilson argues, but clients. Specifications grading takes into account that different clients are going to aim at different outcomes. It makes the expectations clear for each tier and lets the client choose which package to pursue. In Nilson’s telling, this allows the teacher to dedicate the most energy to the students most invested in the course by dint of aiming at the top tiers.

This model is tempting given how frustrating it can be to expend disproportionate amounts of energy on reticent students, but it was also the point that left me most uncomfortable with specs grading. One common proposal in the sample syllabuses Nilson provides is setting not only different levels of proficiency, but also different assignments for the different tiers. I incorporated that into one of my sketches above for the final projects, but even there I have been wondering whether the non-project option ought to require an objective test passed at a certain proficiency since under specs grading—something I’m not wild about given that 1) I am skeptical about the value of such objective tests, period; 2) writing such a test would hand back some of the savings in time; 3) keeping track of who is doing what sounds like a lot of bookkeeping.

However, my discomfort with the different assignments for different levels stems is also philosophical. That is, it feels to me like saving time and becoming a better teacher for the invested students involves allowing students aiming at a “C” to fall behind. The counter, I think, is that this is in fact the point. The way I imagine this grading scheme working in my classes, those students would still be expected to attend and complete assignments for the whole semester and gives anyone who wants it the opportunity to achieve every objective. But if students are not interested, then it empowers them to put their energies elsewhere (courses, hobbies, work, whatever). In other words, the client model simple acknowledges the reality that teachers cannot force people to learn anything they don’t want to learn, particularly at the busiest time of the semester.

I have been thinking about the process as setting two different benchmarks: the “C” level for minimum objectives and the level of proficiency for complex objectives where “A” reaches it in every category and “B” reaches it in some. Specs grading dispenses with the murky ambiguity of partial credit where the “C” student allegedly achieved 75% of a given course objective. Thus, it isn’t the “C” student doing less work so much as they hit one set of objectives, while I am vouching that the “A” student has completed more and more complex work that allows me to certify that they have reached proficiency in the others—I can hope the “C” student developed in these other categories, but the grade makes no claim that they did so.

At this point I am ready to dive into specs grading head first, but I’m also sure that whatever system I come up with in the abstract will require adjustment once I get into a semester. So here’s the question for those of you who have used specs grading: what should I be on the lookout for? Is there anything I’m missing?

ΔΔΔ

I keep a list of pedagogy resources along with links to write-ups I have done on this blog.