Specifications Grading Redux

An alternative grading system. Image: Sage Ross, via Flickr, h/t Robert Talbert.

Last December, I wrote about specifications grading, an idea I first saw on Robert Talbert’s blog Casting out Nines (Co9s is ending, so you can find new posts at rtalbert.org) and wanted to try out in my class. Talbert has blogged about his experiences using the system a few times; you can read some of his specs grading posts here and the rest here. You can also listen to a Teaching in Higher Ed podcast of Nilson discussing the system. Now that the semester’s over, I wanted to let you know how it went for me as well.

First, a word about the book, Specifications Grading by Linda Nilson. I decided to read it rather than just base my grading scheme on Talbert’s and others’ blog posts about it. I hoped that it would give me some nuts and bolts advice for how to make it work in my class. It did, but I almost put the book down after reading this on page six:

What the faculty reap for their endless hours of grading are more grading protests and conflicts with students than ever before. Of course, the reasons behind this student behavior lie largely in the values and beliefs of the Millennial generation and their parents, such as their consumer attitude toward higher education, their distaste for academic learning and the life of the mind, their alienation from standard teaching methods, and their sense of entitlement to high grades in light of high tuition costs.

I am a Millennial, albeit on the old side, and this broad generalization about my generation rubbed me the wrong way. Aside from the fact that it offended me personally, I had serious qualms about taking classroom management advice from someone with such a negative, adversarial view of students and other people my age, especially someone who would say we have a “distaste for the life of the mind.” I do see some amount of entitlement and a consumer attitude, but I have had almost no students who were not interested in learning something. I’m also not at all convinced that Millennials have these traits to a larger extent than members of any other generations. I mention all this here because I hope that Nilson and other educators with this viewpoint will reconsider the way they think of their students (and some of their colleagues). 

That said, I do think the book has good advice overall, and I like the idea that employing specs grading will make assessment a less adversarial process. I used specifications grading for my math history class this past semester, and it was a mixed success.

Some background on my class: this course gives students a writing credit, so the majority of assessment was based on written work. I had 43 students registered on the first day of class, a slight increase over last semester, and my attrition rate was significantly lower; by the end of the course, I had about 40% more papers to grade than I did last semester. Some of the problems I had implementing my grading scheme and with the semester in general were related to the unexpectedly large size of the class. A writing class with 40 students in it is, at least for me, nearly untenable.

Another complication not related to specifications grading was that I had students turn in most of their work on the course website powered by Canvas and used the grading tool there for the first time. I had two problems: Canvas inexplicably lacks a way to save feedback while you’re writing it but before you want to give it to your student, and it logs you out without saving after a certain amount of time. A few times, I was grading during office hours and had a student visit. By the time I was done with the student, the feedback I was working on was gone. The other problem was that some students could not access the feedback on Canvas. This seemed to be a browser issue on their end, not something I could control, but it threw a serious wrench into getting feedback to them in a timely fashion, especially because they sometimes didn’t realize there was supposed to be more feedback.

I used a blended grading scheme, which Nilson calls a synthetic option in the book. The two major projects (a group project and a research paper) were graded using a fairly traditional writing rubric that produced numerical grades. The average of those two grades determined the final grade with additional requirements for each letter grade. These included a certain number of math homework assignments and blog posts that were graded complete or incomplete, as well as turning in paper and project proposals and a paper draft on time.

I allowed resubmissions of one of the homework assignments because a number of students had more trouble than expected on it, but the rest of the homework assignments were one-shot deals. Students could revise blog posts until they were passing. I tried to have a rule that the next revision had to be finished within a week of my giving feedback on the last revision, but I had trouble enforcing that deadline (see below).

I used Nilson’s suggestion of tokens/free passes to turn in assignments late in the hopes that it would cut down on students begging for extensions. I gave each student a total of three for the semester, and there was no bonus for unused tokens. For most assignments, a pass was a one-class extension, and for blog posts, it was a one-week extension. In general, I think the passes helped, but they didn’t eliminate extension discussions altogether. I don’t know if I just didn’t give students enough free passes or if they should have been able to earn more or if I just need to be less of a pushover.

Because I required assignments to be turned in by a certain date for students to get a certain grade in the course, some assignments became high-stakes. If a blog post wasn’t on time, a student couldn’t get an A; if the rough draft of the final paper wasn’t on time, they couldn’t get a B. I am not used to classes like this where a student can’t do more work to make up for a bad week later, and for that reason I sometimes had trouble enforcing deadlines.

I had trouble giving homework assignments an incomplete grade when they weren’t quite up to the level I wanted but reflected a significant amount of effort. Perhaps next time I could borrow Talbert’s idea of Mastery-Progressing-Novice and allow progressing students to resubmit the work. In my case, I would probably use different labels. Something like Passing-Effort-No effort would better reflect the fact that I want students who try the homework in good faith to get a chance to learn from their mistakes.

In reflecting on the deadline and homework complete/incomplete problems I had, I wonder if part of what made them difficult for me is that in traditionally graded courses, you don’t immediately see the consequences of your actions in your final grade. In reality, people who didn’t turn in the first blog post in last semester’s class were unlikely to get an A, but they didn’t immediately know this. Likewise, people who consistently turned in late assignments or homework that was not high quality would get lower final grades, but the effect didn’t happen immediately.

One of the benefits of specs grading that Nilson mentions is that students can to some extent choose what grade they want. Most of my students were shooting for an A, but some only wanted a B or C and chose the assignments they did accordingly. In the future, I like the idea of having students signal to me what grade they would like in the class so I won’t worry about the students who don’t turn in some early assignments.

On the other hand, sometimes a student did all the complete/incomplete work to get a grade that was higher than the grade their paper and project earned; they did all the assignments required to get an A, but their major assignments were B material. Some of these students felt like the rules had been changed on them, and I wish they hadn’t felt that way. One unfortunate aspect of the schedule of the course was that the traditionally graded work came towards the end of the semester, so students were sometimes unsure of where they stood.

I had students turn in rough drafts of their papers about two weeks before the papers were due. I gave a lot of written feedback on the papers, but I did not return them with a rubric that showed what grade they would have gotten if they had been turned in as final versions. I am sure this would have helped my students, and I will try to do that if I teach a course like this again. I also think Talbert is right on the money when he says that giving students more examples of passing and non-passing work would be helpful.

Another overarching problem was that I don’t think students really bought into the specs grading idea. If I use it again, I might spend some time in class or in assigned reading explicitly talking about how the system works and the advantages it has over traditional points grading. I had a syllabus quiz for my students at the beginning of the semester to make sure they understood how the grading scheme worked, but I think I could have made assessment feel more like a partnership if I had let them know exactly why I thought specs grading would be better for both them and me.

I was curious about how specs grading would affect the final grades in my class. I think they were a bit higher than last semester. Most of this I attribute to students (and myself) having a clearer idea of what work needed to be turned in at what time in order to get the grade they wanted, so there were fewer students who thought they could scramble at the end to make things up. For example, they couldn’t get an A or B without turning in a rough draft of the final paper on time. I required a rough draft last semester as well, but I deducted points from the assignment if they didn’t have one. This semester, I got rough drafts from more students, and I think that improved their final papers substantially. I was somewhat torn by the rough draft requirement; it hurt two students who wrote good papers but hadn’t turned in a draft, but I think it helped more students than it hurt.

If you’ve used specs grading in a math class, I’d be curious to know how it went for you.