The Dangers of Criteria-Based Grading

This year, I taught at a charter school that uses criteria-based grading.  To describe the system briefly, the learning objectives of each subject are broken down into individual criteria, called benchmarks.  For each assessment (test, assignment, etc.), every question or item on a rubric is tied to a benchmark.  The student receives a separate grade of 1 (low fail), 2 (high fail), 3 (proficient), or 4 (advanced) for each separate benchmark.  The student’s grade in the course is the percentage of benchmarks the student has passed (with a grade of 3.0 or better).  A grade of 70% or higher is passing.

This sounds great in theory.  Students know exactly which skills they need to master.  Teachers know exactly which skills they need to teach, and how their students are performing on each one.  Administrators can use students’ performance on each of the criteria as a measure of teacher performance.  For courses that have state frameworks, the criteria are aligned to the frameworks and constitute evidence that the teachers are teaching to the frameworks.

Unfortunately, what appears to be getting lost is students’ ability to aggregate their knowledge, think flexibly, and solve problems that are different (sometimes only subtly) from the ones they have already seen.  I taught three different science courses this year–chemistry (mostly 10th graders), honors physics (mostly 11th graders), and environmental science (mostly 12th graders).  Most of these students have been in the school since 6th or 7th grade, meaning that they had been in the school for 3-6 years by the time I taught them.  I saw the same trend across all three classes.

The school does not mandate direct instruction (DI), but DI accounts for much of the teaching in the school, which I think is one of the factors that contributes to the problem.  DI is well suited to assessments on individual criteria–students are taught how to solve the specific types of problems that the criteria describe, and are then assessed on their ability to solve those problems on their own.  When students pass the assessment, the teacher moves on to the next set of criteria.

In my own teaching, I try to focus on high-level thinking, with plenty of multi-level problems that combine concepts.  If multiple types of problems involve some of the same quantities, I give students problems that combine them.  (I once gave an AP chemistry summer assignment problem that took this concept to the extreme–it involved a limiting reagent acid-base redox reaction in which one of the products was a radioactive isotope of a diatomic gas that had a half-life of 37 minutes and decayed into a monatomic gas.  Students were asked to find the pressure in the tank after a specified amount of time.)

This year, in all of my classes, most of my students were completely stymied by even the simplest combinations of concepts or the smallest deviations from problems as presented in class.  When I would ask them to solve a problem that was even a little different from what they had seen in class, many of them would stare at the blank piece of paper and do nothing.  When I prompted them, they replied simply that they didn’t know what to do.  None of them attempted to take out their class notes or periodic tables to try to figure anything out; they simply waited for me to give them instructions for how to solve that particular problem.  This scenario repeated itself over and over again throughout the school year.

Interestingly, I had raised this concern during the new (to the school) teacher orientation before the school year started.  The head of school had asked us to discuss some of the advantages of the criteria-based system over a traditional grading system.  I offered that the traditional grading system specified only the big picture.  The teacher had to be trusted to distill the big picture down to its components, but there were no checks in place to ensure that the teacher did so.  However, I went on to point out that the criteria-based system specified the components.  The teacher had to be trusted to aggregate them, but that there appeared to be no checks in place to ensure that the teacher did so.  That comment turned out to be prophetic.

The other significant problem with this school’s implementation of criteria-based grading was that there was no provision for weighting of assessments.  I regularly give homework assignments, “open friend quizzes” (a type of formative assessment in which students are allowed to work with a partner), lab write-ups (in a lab notebook), formal lab reports, and pencil-and-paper tests.  The same criterion might be assessed with different levels of difficulty on a homework assignment, on an “open friend quiz”, as part of a lab write-up, on a unit test, and on the final exam.  For my part, a more thorough and complete assessment of a criterion needs to be weighted more heavily than a less thorough and complete assessment.  Otherwise, I am stuck with either being unable to give a student a 4 for thorough and complete data analysis on a lab write-up if the calculations did not present enough of a challenge, or with giving the student a 4 for analysis on the lab, but then watching the student earn a 2 for the same criterion on the unit test because the problem on the test was more challenging.

At this point, I have a case study in how one particular criteria-based grading system can fail.  I’m not to the point of concluding that it’s an unworkable idea, but I would very much like to see a successful implementation of criteria-based grading that preserves the notion of teaching students to combine different concepts that were assessed under different criteria, and that also successfully integrates a weighting system that takes the difficulty of the assessment into account.

+5

-5