Specifications-Based Grading for Assessing Student Learning

Kristi Closser, California State University-Fresno

What is the difference between a student who receives a 79% in a course and one who gets 80%? At my university, we do not have +/- designations associated with grades, so if I were following a typical scale, the first student would receive a C and the second would get a B. Then what is the meaning of assigning students a letter grade such as a B or C? In the era of grade inflation, does a B indicate average? Above average? From my very first semester of having to assign grades, I detested trying to make meaning out of percentages.

Last semester I was introduced to specifications or mastery based grading by one of my colleagues. While this idea is not particularly new, it is also not a widespread practice—especially at the university level. As I've only done this for one semester, I do not pretend to be an expert, and there are better resources available from real experts, some of which I include as references at the end of this essay. So here I will focus on what I know, which is my own experience with this assessment practice.

The course I taught using this format, was actually an organic chemistry course. There are many variations for undertaking specifications-based grading; for my course, there were a total of 16 learning target assessments (LTAs), which were short quizzes designed to address specific course objectives. The first 7 of these were considered core ("atomic-tier") topics that all students had to pass in order to pass the course, the next 6 were "molecular-tier" topics that were all important to know for being successful in subsequent courses, and the final 3 were "supramolecular-tier" topics which did not introduce new material, but combined earlier material. Each LTA was graded satisfactory/unsatisfactory and final grades were primarily based on the number of LTAs that were satisfactorily passed (final grades also accounted for a cumulative final and participation, but these never lowered the grade from what students earned on the LTAs). For an A students had to pass 14 LTAs, 11 for a B, 9 for a C and 7 for a D. One of the key differences in this system vs. standard grading systems is that students were allowed to retake LTAs which were unsatisfactory. As this was my first time using the system there were some issues with multiple retakes that I will address next time, but overall the system worked pretty well. The final letter grade distribution for the class was very similar to what is obtained using a standard point based system, and completely bypassed the headache of trying to scale the final or adjust grade cutoffs at the end of the semester.

Although organic chemistry does not involve a significant amount of computation, it shares many of the high-anxiety/high-stakes traits of courses which do focus on developing computational skills. One of the most powerful aspects of allowing a quiz retake or more generally allowing students to redo an assignment is that it removes some of the anxiety students have about the course and refocuses the students on learning the material rather than getting a specific grade. If a student receives an "unsatisfactory" mark, it does not mean that they missed the chance to learn the subject, but merely that they have not mastered it yet. Although I only have anecdotal data at this point, I believe that it also significantly helps level the field for students who begin the class at a disadvantage due to prior preparation. It gives these students a chance to see where they miss questions and then after spending a bit more time with the material demonstrate that they have learned the new topics.

Additionally, using a mastery-based grading system had the effect of completely transforming my office hours. Rather than having students come to ask how they could get 2 more points, I had students coming to go over specific topics (and I had three times as many students coming to ask for help from my specifications-based class than from the class I graded using a traditional system). Students also felt more in control of their grades, and the focus of the class was shifted towards learning to enable earning a particular grade.

In this upcoming semester I will be applying a similar grading scheme to my physical chemistry course, which does include more computational work. In many ways the specifications or mastery based grading is an excellent fit for this type of class. In this class my students will have a series of quizzes, and also more involved problems that need to be solved. These will all be graded using an EMRN scale: E= exceptional, M=meets expectations, R=revision required, and N=not assessable. While this is slightly more nuanced than a simple pass/fail, it gives students a bit more feedback on their performance. Another change I will be making is that students will have no more than 2 written attempts on the quizzes, any subsequent attempts will have to be done as an oral exam during office hours. I am hopeful that the threat of an oral exam will be enough to make sure they take the written attempts more seriously than students did this last semester. Additionally, students will have to spend an increasing amount of "p-chem pounds" for retaking quizzes or assignments. Everyone starts the semester with 2 £, and more can be earned by completing designated problem sets. The EMRN scale also means that for assignments, students can resubmit freely if they earn an R, but must spend a pound if the grade is an N. Clear expectations will be set for each assignment or quiz for what constitutes each level.

Assessment of student work can be simplified by eliminating partial credit on assignments and by giving students a chance to earn the grades they want. By changing the assessment system, retaking quizzes and reattempting assignments can be used to give students a chance to focus more on the process of learning. I'm moving away from points-based grading systems, and don't intend to ever look back.

References and Resources
Nilson, Linda B.; Specifications Grading: Restoring Rigor, Motivating Students, and Saving Faculty Time; Sterling, VA: Stylus Publishing, 2015.
Stutzman, Rodney Y.; Race, Kimberly H. "EMRF: Everyday Rubric Grading", Mathematics Teacher, 97, p. 34-39, 2004.
Talbert, Robert; Blog—currently under migration (8-14-18); some older posts linked below http://www.chronicle.com/blognetwork/castingoutnines/2014/12/22/an-update-on-the-specifications-grading-process/; http://www.chronicle.com/blognetwork/castingoutnines/2015/01/19/putting-specs-grading-to-work/; http://www.chronicle.com/blognetwork/castingoutnines/2015/03/06/observations-and-changes-for-specs-grading-8-weeks-in/
Google+ community: https://plus.google.com/communities/117099673102877564377

Downloadable version of this essay

Specifications-Based Grading for Assessing Student Learning (Acrobat (PDF) 550kB Aug16 18)