If you’ve spoken to a techie or tech bro recently, you’ve probably heard the phrase “fail fast, fail often.” The idea is simple: value trying and learning from mistakes, rather than going slow and demanding perfection. I teach this in my classes, especially in engineering classes, but also in the more traditional science classes where I want the kids to feel free to try and make mistakes in order to grow.

As a teacher, when I hear “failure,” I think of the number 60 percent. For students, a bad grade is closely linked to the stress of not meeting a required standard of performance. My school has seen a lot of grading and failure over the past year. Like many other schools, we’ve debated grade inflation and the merits of flexible grading policies that many people have adopted for pandemic-era learning.

In my ten years in schools, I have never seen a debate where teachers (and students) were so deeply divided about the right way to proceed. At times, discussions about these grading policies felt like people cheering on competing baseball teams: team grading based on championship versus team grading the old-fashioned way. “Liz is against repetition and she will argue with you,” is one sentiment I’ve heard in meetings.

Earlier this year, I resolved this conflict in my own classroom when I began co-teaching a course called “Design at Human Scale” with my colleague Brendan. As a math teacher, Brendan has played a large role in leading our school’s math department’s competency-based grading. He assigns grades to students based on benchmark topics that they are allowed to repeat until they demonstrate the required proficiency. He uses a 4-point scale that he adjusts to reflect the 0-100 percent used schoolwide.

I, on the other hand, remained much more traditional. I gave assessments at regular intervals, grading them based on the percentage of correct answers, and then adding up all of the assessments to give a final score. I wasn’t sure how we would consolidate these assessment philosophies.

I had learned enough about grading to have some doubts about my system, but I was still tied to “normal grading.” The idea of ​​abandoning my policies seemed unimaginable. I felt that my grading made sense, and people who disagreed with it just didn’t get it.

However, I found it difficult to give grades that seemed subjective. This subjectivity was common in my engineering class, which had more creative tasks than a physics class.

My engineering classes are all about iteration, where projects start small and simple and get more complex. Failed prototypes are good things to grow from. In my fidget spinner project, we start by making paper fidget spinners with spaghetti bearings. A next version will be made out of cardboard to test size and shape before a final, laser-cut product. But I graded the project based on a final paper that didn’t reflect the value of iteration.

Given my difficulty with subjectivity, I strove for better, more comfortable grades, and suddenly engineering was an “easy A.” (Any university-level engineering student would be surprised by this, given how challenging engineering has traditionally been.)

I was curious if my grading would reflect the “fail fast, fail often” mentality I was teaching. My new design class offered me the opportunity to try.

When planning the course, Brendan and I worked together to design a system that we were both interested in trying out. We considered our course objectives and identified four key learning focuses: design, iteration, tools and software, and community.

For each project we undertook, we asked students to submit evidence of what they had learned in each category. We reviewed the evidence and determined whether the student’s work showed progress and passed. If the evidence was not sufficient, students had the opportunity to resubmit it.

This meant that we never graded anything except the students’ reports of their own development. I didn’t have to look at every clock the students designed and give them a grade like a 90 or an 88. Instead, the student took over the work and told me what they learned from making the clock. Brendan and I used a table that correlates the number of pieces of evidence with a grading scale to convert that evidence into a numerical grade.

While our system still needs improvement, I like it enough to keep tinkering with it and try to apply it elsewhere in the future. The moments of subjectivity have been replaced with a more objective scale. Creative work should not be graded on whether I am happy with it, but on how well students justify their design decisions. It’s still not perfect. But it doesn’t have to be. I just had to be willing to “fail fast and often” myself.

Co-teaching this course has helped me learn how I want my students to learn: trying new things, failing, and learning from that failure. Faced with differing professional opinions on grading, I’ve learned that we succeed when we’re open-minded and curious, rather than argumentative. There is no perfect, easy, or obvious solution to grading well.

Being more open to iteration in my own practice has led to greater professional growth than all the debates about the merits of repetition, correction, and mastery. I am happy to have found the learner within myself that I want my students to be.