team sits silently, and Kaori, not expecting an answer, continues, “It’s important to start with understanding why we are making this change. Our mission is to ensure the most accurate and clear communication about learning to promote success for all students. Evidence-based grading principles support this mission.”
John then takes his turn. “It’s frustrating, but our current system, which we’ve been using for decades, doesn’t support our mission of clear communication. This came to me when I thought about real-life student experiences with grading and reporting practices. Let’s start with a scenario. Let’s suppose a student gets the following grades on five exams for one six-week grading period: 40, 60, 80, 90, 90. What grade does the student deserve?”
Mario replies, “I know this is not the answer we will give by the end of this meeting, but I would say 72 percent based on the way we calculate grades now.”
“Well,” John says, “a few things come to mind when I hear that. First, the student’s last two assessments yielded a score of 90, and also the student never scored in the 70s at any point during these assessments. If we look through the student’s grades, it appears that over time, he made significant improvements. The student learned over time. If we calculate the student’s grade average, aren’t we really discounting his growth? Aren’t we assigning a grade for the student based on what he wasn’t able to prove instead of what he is now able to prove?”
Maya speaks up. “I agree, but the student also has to be accountable for past mistakes. Averaging all those grades together is more accurate because the student was only at 90 percent for a short time period out of all the assessments. Therefore, a 72.5 percent, or C grade, is really a good picture of the student during the course. He didn’t do well the whole time, just part of the time.”
John says, “That may be so, but think of it this way. Do you remember learning how to ride a bike? When you learned, did you factor in all the times you fell to determine whether you could ride, or did you just finally learn to ride? You didn’t just average your ability to ride the bike and say, ‘I ride a bike at 72 percent.’”
Team members nod in agreement. John continues, “Would you consider a student who can now fluently speak another language not fluent because she made many mistakes along the way? Of course not. Or how about a student who didn’t know algebra or chemistry at the beginning of the school year but learned it by the end?”
John moves to make his point. “All learning is based on growth. In fact, that is the definition of learning. Evidence-based grading is a growth-based learning model and supports the expression of skill acquisition and knowledge. Our current system of grading does not express anything but percentages or point earnings. It doesn’t communicate learning’s growth and development.
“Let’s consider another example. Suppose a student gets the following scores: 0, 0, 0, 100, 100, 100. What percentage will she receive?”
Kevin says, “Traditionally, the student would get a 50 percent. She would fail the class.”
“Correct,” John replies. “Now, how many 100 percent grades would the student need to get in order to offset all those zeros and earn an A?”
The group is slow to answer this time. Some members mumble a few answers, but nothing seems correct. John explains, “In order to get an A grade, the student would need twenty-seven more 100 percent grades in the gradebook to offset the initial three zeros. In other words, it is almost impossible to outpace a particularly low grade, especially a zero.”
“So, you are saying we shouldn’t use zeros. I get it. But what about the student who just doesn’t do the work?” asks Maya.
Kaori says, “Behavior and academics must not coexist in one single letter grade. The comingling of behavior, skills, attendance, attitude, work ethic, and skills performance creates lack of clarity about why the student gets a certain grade. We then assume a lot about what is behind the grade. Think of the letter grade B. Some parents may think it means smart but the class is hard, while other parents may think not as smart as other students in the class. Some may think that their child didn’t work hard enough. Too many assumptions cause the grade to be less accurate. Therefore, it is not aligned with our mission, which is ensuring the most accurate assessment and communication of student growth and performance.”
John says, “We see this same problem in the assessments themselves. Let’s assume you have an exam this week. One student skips the test and another gets all the questions wrong … yet they both get a zero in the gradebook. So, what does that zero represent? Does it represent a lack of effort or a lack of knowledge? You really don’t know the grade’s intent on a report card.”
The team acknowledges his point. John continues, “So, can we agree now that zeros are not effective, averages do not work, and grade information must be reported separately from behavior as a way to communicate meaningfully?”
The team understands exactly what John is saying, and each member thinks about how the years of past grading practices might not have been equitable to students.
Kaori starts the next segment. “In evidence-based grading, we assess students on a gradation of learning that has four levels. Why only four levels? Let me explain. It is essential as assessors that we possess the capability to articulate a clear description of each level of achievement as well as the differences between these levels. This clarity about each level of achievement is not only important for equitable and just assessment of student performance, it is also important for feedback, curriculum, and instruction. In fact, evidence-based systems are based on this gradation of achievement and the ability to articulate it. Without clarity, it is impossible to assess students accurately.
“There are four levels in our system,” Kaori continues. “Fewer levels means that students are classified more accurately. As Thomas R. Guskey points out on page 36 of his 2015 book On Your Mark, ‘essentially, as the number of grade categories goes up, the chance of two equally competent judges assigning exactly the same grade to the same sample of a student’s work diminishes significantly.’ Let’s do a little exercise.”
“We currently use a one-hundred-point scale,” John says. “Together as a group, think about an assignment you recently gave to your students.”
Joni says, “A free-response writing assignment about the Battle of the Bulge.”
“Great!” says John. “Now think about the students’ grades, and ponder this question: What is the difference between a student who gets an 85 percent and a student who gets an 86 percent? Keep in mind that in order to have an accurate and fair grading system, the assessor must be able to articulate the difference between these two percentages.”
The team is silent for a few seconds, and then Mario laughs. “I can’t,” he says. “There is no real difference between 85 percent and 86 percent.” The others agree.
“The more levels we have,” John says, “the more we run the risk of potentially giving an incorrect rating. Even worse, we give inaccurate feedback. This is why we must have the fewest achievement levels possible that still promote quality feedback. This is why we use four.”
Kaori says, “Evidence-based grading is based on achieving a level of proficiency—proficiency in a skill or proficiency in consolidating information into actionable thoughts. This proficiency is assessed by a gradation of achievement that represents an assessor’s expectations. Expectations have gradations, and you use them to evaluate the current evidence of performance. Does that make sense?”
Again,
