Ungrading Has an Equity-Related Achilles Heel
Implicit bias is unavoidable in ungraded courses
Welcome to a new semester of Grading for Growth! This semester, we’re beginning with a guest post by Jayme Dyer, PhD. Jayme is an Adjunct Instructor of Biology at Durham Technical Community College in Durham, NC where she teaches General Biology. Her class sizes range from 20-45 and her students are diverse in age, academic background, career goals, religion, disability status, gender, and race/ethnicity. Approximately 30% of her students are white, 30% are Black, and 30% are Hispanic. Jayme is white woman from a low-income and foster care background.
Many instructors who are interested in helping their students learn more, and learn in more authentic ways, have adopted the pedagogical approach of ungrading. While instructors use the term ‘ungrading’ to mean different things, I refer to ‘ungrading’ here as the instructor’s choice to reduce, or make absent, points and assignment grades in an effort to focus student attention on instructor feedback and personal growth. There are many compelling arguments – primarily from personal testimonies – to the power of ungrading as an approach to improve learning, develop students’ autonomy, and make teaching and grading a more fulfilling experience for both students and instructors.
But, ungrading has an equity-related Achilles heel.
Namely, for courses that require a grade at the end of the semester, the allocation of grades in an ungraded course is highly susceptible to the instructor’s - and also sometimes the student’s - implicit biases. These biases can perpetuate systemic inequities, including in students’ access to privileged academic opportunities well after the course has finished.
I have implemented ungrading in multiple semesters, in multiple formats (to varying degrees of success) in my community college General Biology courses. But I won’t be using it again. In this post, I explain why.
Final Course Grades Matter
Final course grades can have significant impacts on students’ lives after a course is finished. Students often have to earn a minimum GPA to maintain access to financial aid. Acceptance to privileged academic opportunities – like summer research programs, honors programs, and graduate school, among others – often use GPA as a significant metric to determine acceptance. A high GPA can even result in lower car insurance payments.
Additionally, grades matter in a race-dependent way. It has been documented for decades that PEERS (People Excluded based on their Ethnicity or Race) are disproportionately lost from STEM programs relative to non-PEERs (Asai, 2020). A 2022 study of 110,000 undergraduate student records from 6 large research-intensive universities showed that “even after controlling for academic preparation in high school and intent to obtain a STEM degree… the association between [low grades] in an introductory STEM class and failure to obtain a STEM degree is stronger for under-represented minority students than for other students” (Hatfield et al. 2022). In other words, low grades have a disproportionate effect on PEERs in STEM courses.
Of course, there are exceptions. As a PhD student, I was told explicitly that my grades in graduate school didn’t matter. And they didn’t: for the awards I won and the positions I pursued after graduate school, no one ever looked at my grades. But for most of our undergraduate students, final course grades matter, and they matter in real, lasting ways. The fact that grades have a greater impact on PEERs than non-PEERs means it is critical that we examine whether grade allocation is equitable.
By “equitable grade allocation” I mean that the same evidence of learning should earn the same grade regardless of a student’s identities. This is different from “equitable access to learning,” which focuses on the ways that students with some identities experience unique barriers to learning. I absolutely think we should design grading systems to improve equity in learning, but in this post I’m attending only to the steps after learning: does the same performance earn the same grade for all students? In an equitable grading system, it should.
How Grades are Allocated in Ungraded Courses
Most instructors who implement ungrading are required to submit a grade at the end of the semester. Without assignment grades or points to add up from throughout the semester, how do instructors decide on a final course grade? There are essentially two ways to do this: 1) The instructor holistically reviews a student’s body of work from throughout the semester and makes a call about the appropriate final grade, or 2) The instructor and student collaboratively determine the student’s final grade (the basis for many instructors calling the practice “collaborative grading” rather than “ungrading”).
In a collaboratively-graded course, the instructor and student determine the final course grade together. Typically, at the end of the semester the student submits a portfolio of their work, often combined with a scaffolded reflection in which the student self-assesses their growth and proposes a final course grade. Then, together in a one-on-one meeting, the instructor accepts or proposes a different grade (either up or down) from what the student proposed. When the instructor and student disagree, they engage in a discussion where they each justify their proposed grade, before agreeing on a final course grade together.
The benefits of collaborative grading include giving students an opportunity to explain unique circumstances that may shed light on certain aspects of their performance. In a collaborative grading conversation I had with a student, he explained that for one unusually low-performance quiz, he had simply forgotten to turn over the paper and therefore only completed half of the quiz. In another meeting, a student explained that partway through the semester she realized that she didn’t need to do the pre-class homeworks to do well on the assessments, so she stopped doing them. Thus, by taking a holistic approach to determining the final course grade, students’ foibles, choices, and life circumstances can be taken into account, rather than counting against them de facto.
However, there is an inherent problem to determining final course grades holistically. Whether final course grades are determined collaboratively, or solely by the instructor, they are both liable to be influenced by implicit bias.
Ungrading and Implicit Bias
Extensive research has shown that when instructors know the identity (or presumed identity) of the student who submitted an assignment, they grade otherwise equal performance more harshly for certain identities. For example, an essay that is associated with a woman’s name is graded more harshly than the same essay associated with a male name or an “anonymous” label (Jackson, 2016).
According to the American Psychological Association, implicit bias is a “a negative attitude, of which one is not consciously aware, against a specific social group.” In the example above, graders display implicit bias based on perceived gender. But implicit bias is not just about gender; a student’s race or ethnicity, migrant status, and physical attractiveness have all been shown to impact how a grader rates student performance (see the meta-analysis by Malouff and Thorsteinsson, 2016).
Implicit bias is, by its very nature, not something instructors are actively aware of, but it can affect a student’s grade nonetheless. Implicit bias cannot be erased, or consciously changed. While targeted trainings and intentional mindset approaches can reduce implicit bias, we all have it (I do! you do!), and the equitable – the ethical – thing to do is avoid situations where your implicit bias can “act out” in the world by perpetuating systemic inequities.
Ungrading is exactly that: a situation where the instructor’s implicit bias can perpetuate systemic inequities. To be clear, even in a traditionally graded class, implicit bias can perpetuate inequities when the instructor makes a call about an assignment grade, but the scale of the impact is much bigger when the call being made is the final course grade. And for collaboratively-graded classes, the impact of implicit bias is even worse because the student’s implicit bias can also perpetuate systemic inequities.
Before I met with my General Biology II student Nida (pseudonym) for our end of semester one-on-one collaborative grading meeting, I read her portfolio and reflection. At first, I didn’t look at her proposed grade. Based on her performance on assessments and her engagement in laboratory and homework activities, I genuinely couldn’t decide if she should earn an A or B.1 Then I looked at the grade she proposed for herself: B. In our one-on-one meeting, I asked her why she thought she deserved a B instead of an A. Her answer boiled down to, “I’m a B student.” I was already on the fence about her grade, so I let her convince me to give her a B. Nida is a Black woman.
I’ve had other students push back when I propose a grade lower than what they proposed for themselves. This has happened rarely, partly because I usually don’t bump “down” grades during collaborative grading, but more importantly, because most students don’t feel comfortable arguing with a professor. While I love when my students disagree with me (and use evidence to do it!), I’ve noticed that the only students who have done this with me are men.
Over the years, as I’ve sat through one-on-one collaborative grading meetings, I’ve grown more uncomfortable. Which students are more likely to propose a higher grade for themselves? (Not women, in my experience). Am I as likely to “bump up” a Black student’s proposed grade as a white student’s? Am I as likely to accept a grade-related argument from a Hispanic student as from a white student? Is it equitable to let the students who are the most confident in arguing with a professor push their grade up? If Nida had been white, or male, would she have been more likely to propose an A instead of a B?
The potential for implicit bias to affect the allocation of grades makes me deeply uncomfortable with ungrading. (I’m not the only one.) Whether the final course grade is determined collaboratively, or by the instructor alone, there is substantial potential for implicit bias to “act out” by perpetuating systemic inequities.2 Of course, implicit bias can “act out” any time an assignment is graded non-anonymously, but the size of the equity impact is that much bigger when the judgment call is about a student’s entire course grade.
When a final course grade is required, there is no way for ungraded courses to avoid the implict-bias-prone judgment call; the process is baked into the system! That is the equity-related Achilles Heel of ungrading.
Remember, final course grades matter. To allocate grades in a way that is deeply influenceable by implicit bias is to perpetuate systemic inequities. I cannot and will not do it any longer.
What I Do Instead
Let me be clear: I’m not advocating for traditional grading instead of ungrading. Many aspects of traditional grading include both implicit bias or punitive policies that disproportionately harm our most vulnerable students. If you are early in your journey of thinking about how grading impacts classroom equity, I highly recommend reading the book Grading for Equity by Joe Feldman.
Here is a brief overview of what I do in my classroom to support student learning and improve equity: Every assessment in my course has reassessment without penalty built-in, either through a “retry” opportunity (i.e. the final exam covers the same content as weekly quizzes) or a built-in revision option. To support students who demonstrate mastery on the first or second assessment attempt, I use a points-based grading policy called Multiple Grading Schemes3; in my courses, one grading scheme weights quizzes (first attempt) more heavily while another grading scheme weights the final exam (second attempt) more heavily. I assign frequent pre-class homework assignments to scaffold student practice throughout the semester; because completion of these homework assignments is a measure of compliance and not learning, they are graded only on completion, and some grading schemes don’t count this component in the final grade at all. At the end of the semester, I calculate each student’s grade using each grading scheme and they get the highest grade.
Preliminary data from nearly 1500 students in Biology, Math, and Physics courses at my community college that have used Multiple Grading Schemes show that Black and Hispanic students are more likely than white students to earn a higher grade because we used Multiple Grading Schemes4. At my institution, Black and Hispanic students, on average, earn lower grades in STEM courses. Thus, implementing a grading policy that is equally applied to all students but that disproportionately benefits Black and Hispanic students is one step closer to equitable grade allocation.
Is Ungrading Redeemable?
There isn’t much research yet on ungrading’s impact on student learning (specifically, removing or reducing points and assignment grades in a course)5, but lots of anecdotal evidence suggests that it can, at best, have a transformational effect on the learning process. Many instructors report how removing points and assignment grades frees students to focus on feedback and learning instead of grades, and it facilitates students taking ownership of their learning process. I do think that ungrading, for some students6, can promote more and better learning.
In contexts where grades don’t matter (like my PhD program), or where there is no requirement to submit a final course grade, I think ungrading can be a wonderful approach to support student learning.
But as long as final course grades must be assigned, and as long as final course grades are routinely weighed in acceptance decisions for privileged academic opportunities, and as long as we have implicit biases, I worry deeply about the ways in which assigning end of course grades in an ungraded course may be perpetuating systemic inequities. Perhaps the costs from inequitable allocation of grades are outweighed by the boost to learning and student autonomy that comes from ungrading. But without sufficient research, I’m not convinced yet.
If you’d like to read more of Jayme’s thoughts on teaching, subscribe to her blog at https://jaymedyer.substack.com/ .
At my institution, we only assign letter grades, without +/- designations.
The assumption I’m making here is that in an ungraded course, whether the final course grade is determined collaboratively or not, the instructor knows which student is associated with the work. If an instructor could assign final course grades based on identity-free portfolios, that might solve the implicit bias problem, but I think it would be impossible to do in practice.
To be published this year, hopefully!
There is research showing that students are less attentive to written feedback when it is presented alongside a numerical grade, which speaks to a potential benefit of ungrading, but there isn’t, to my knowledge, much research into the impacts of ungrading on learning or on equity.
While a goal of many alternative grading practices is to improve implicit motivation, it's unreasonable -- and practically speaking, impossible -- to pretend that only implicit motivation matters. Because ungrading removes points and grades as behavioral incentives, it may not be the best course structure for students who need or benefit from external motivation, such as students taking a course to fulfill a curriculum requirement or students with ADHD who benefit from external structures.