Undergraduate Education

How an alternative grading system is improving student learning

Interest grows in specifications grading, an approach to assessing students that could offer more transparency and consistency

Celia Henry Arnaud

2021-04-25 22:17:55

 

Katty Huertas

In brief

What’s in a grade? Extensive use of partial credit in points-based grades can make it difficult to tell what a student has actually learned. A small but growing number of chemistry professors are adopting an alternative grading system called specifications—or specs—grading. This approach is based on clearly defined learning outcomes, pass-fail grading with no partial credit, and multiple opportunities for students to demonstrate mastery. Professors who have adopted the approach have seen improvements in student learning. And early indications suggest that specs grading could be more equitable than conventional grading approaches.

For many professors, grading student work is the least enjoyable part of their jobs. “None of us get into teaching to grade,” says Renée Link, a professor of teaching in the Chemistry Department at the University of California, Irvine. She’s uncomfortable with how grades become gatekeeping devices, causing graduate schools and potential employers to discount students’ scientific abilities. Plus, grades aren’t as accurate and precise a measure of student learning as many people assume. On top of that, in large courses like Link’s, grades are routinely normalized to adjust for differences among graders. “It’s just a terrible system,” she says.

To make the experience of grading less onerous for her and her teaching assistants, Link has joined the small but growing number of chemistry professors who have adopted an alternative grading system known as specifications—or specs—grading.

This system eschews the conventional points-based approach to grading. The key components of specs grading are clearly defined learning outcomes (measurable statements of what a student will know or be able to do after taking a course), pass-fail grading of assignments with no partial credit, and multiple opportunities for students to demonstrate mastery of learning outcomes.

But specs grading does more than make grading easier on professors. It makes grading more transparent for students. And it gives professors and students a clear picture of what students have learned and what they still need to work on. Because success is defined by detailed rubrics—the eponymous specifications—grading is more consistent and has the potential to be more equitable.

Nuts and bolts

Joshua Ring of Lenoir-Rhyne University and Santiago Toledo of St. Edward’s University were both early adopters of specs grading. Like many, they were inspired by a 2014 book by Linda B. Nilson, who is now director emerita of the Office of Teaching Effectiveness and Innovation at Clemson University. In her book, Specifications Grading: Restoring Rigor, Motivating Students, and Saving Faculty Time, she outlines the rationale for specs grading—making grading easier for professors, improving rigor in courses, and encouraging students to take responsibility for their learning—and the mechanics of how it works. Her approach is flexible enough that it can be adapted for any course that can be constructed around defined learning outcomes.

Before adopting specs grading in 2016, Toledo had put substantial effort into writing clearly delineated learning objectives for his general chemistry courses (J. Chem. Educ. 2015, DOI: 10.1021/acs.jchemed.5b00184). Once he defined the learning outcomes, redesigning his assessments was a natural next step.


Changing distributions
When Justin Houseknecht of Wittenberg University switched to specifications grading for organic chemistry, a higher percentage of students got As and Bs.


09915-cover-chart.jpg

Source: Justin Houseknecht, Wittenberg University.
a DFWs stands for Ds, Fs, and withdrawals.

Within each outcome, students have to complete multiple tasks, including homework assignments and “checkpoints,” Toledo’s version of quizzes. “We set specifications to what it means to get proficiency on each of those tasks,” he says. “The big picture is that we’re trying to give students feedback according to learning outcomes.”

The number of outcomes varies among professors who use specs grading. Some have as few as a dozen. Others have 30 or more. In organic chemistry, for instance, a student might be expected to master skills like drawing Lewis structures, predicting reactivity, and drawing reaction mechanisms.

Rather than rely on only a few high-stakes exams to gauge student progress, specs grading as used in chemistry involves frequent quizzes, each addressing one of the established outcomes. Quizzes or other assessments are typically graded on a pass-fail basis. Students either get a question right or they don’t—there’s no partial credit. The most widely used threshold for fulfilling an outcome is answering 80% of the questions on a quiz correctly.

The number of learning outcomes mastered determines students’ grades. Some professors divide their outcomes into two buckets: “essential” and “general.” Students must master all the ones designated essential in order to pass the class. Each additional general skill in which they demonstrate proficiency raises their grades.

This system means students know exactly how many outcomes they have to achieve to get the grades they want. “If your goal is a C, here’s what you have to do,” says Mary Beth Anzovino, who uses specs grading in her organic chemistry classes at Georgia Gwinnett College. “You don’t have to pass all the generals, but you do still have to pass all the essentials, because by earning a passing grade in this course, I’m basically signing off on saying this student can do X, Y, Z, A, B, C. Some of the other skills are perhaps less crucial.”

Justin Houseknecht, who uses specs grading in his organic chemistry classes at Wittenberg University, says the distinction between essential and general learning objectives confused some of his colleagues at first. He has 12 essential outcomes and roughly 18 general ones. His colleagues asked, “If something’s not essential, why are you teaching it?” Houseknecht says. He explained to them that the essential concepts underpin the general ones.

Students get multiple chances to pass each outcome. In that sense, specs grading more closely resembles how learning and assessment happen in real life. Mai Yin Tsoi, an organic chemistry professor at Georgia Gwinnett College, says the multiple opportunities are like the tests required to get a driver’s license. Failing the first driving test doesn’t put a license out of reach. Instead, people can study, practice, and retake the test.

By earning a passing grade in this course, I’m basically signing off on saying this student can do X, Y, Z, A, B, C.

Mary Beth Anzovino, assistant professor of chemistry, Georgia Gwinnett College

Most chemistry professors using specs grading allow students a free retake for each outcome. Rather than give students unlimited attempts, many users of specs grading use a token system for earning additional opportunities. For example, in Ring’s organic class at Lenoir-Rhyne, students earn tokens by watching the videos he uses instead of in-class lectures or by completing “bridge” assignments that lead from the videos to material and problems he plans to present in class.

The ability to earn extra retakes provides the students a safety net. But in the process of earning the retakes, the students are engaging with the material and doing the work required to learn it. “Earning the retakes means that you’re probably not going to need them in the first place,” Ring says.

While all those retakes might sound like a scheduling headache to busy professors, schools have gotten creative with sharing the workload. At Georgia Gwinnett, several professors in different departments have adopted specs grading. The faculty members have formed a cooperative through which they each sign up for a time for proctoring retakes. The retakes all happen in a dedicated room in the college’s learning center. Students can sign up for any available time, no matter which professor is proctoring.

“It took a lot of the time burden of actually administering the retakes off of the individual faculty,” Anzovino says. “I could have up to 24 students in that room doing retakes from 10 different classes during my proctoring times.”

Another challenge created by the opportunity for multiple retakes is the need to write multiple versions of quizzes for each learning outcome. To reduce the number of versions they have to write, they don’t return some of the graded quizzes, so they can reuse them in subsequent semesters, Anzovino says. “I never change the quizzes too much because the learning outcomes are so fine grained. It’s a pretty specific set of skills that we want to be able to demonstrate on each of those quizzes.” But creating those quizzes still takes time.

Scaling up

Chemists at liberal arts colleges and universities with smaller class sizes were early adopters of specs grading, but some professors at universities with large classes are experimenting with the approach.

Link coordinates UCI’s organic chemistry lab sequence. Each quarter, many teaching assistants (TAs) teach more than 1,000 students in the course. In her conventional points-based grading system, she would find that some TAs were lenient and others were harsh. As a result, she would need to adjust students’ grades at the end of each term.

“One of the challenges with grad student TAs, especially early on, is that they still identify more as a student than as a teacher. They’re just like, ‘Well, you tried hard; let me give you credit,’ ” Link says. “I understand wanting to give them credit for their effort, but you’re actually hurting them because you’re making them think they know something when they don’t.”


Try, try again
Specifications grading gives students several chances to master a skill, such as naming alkanes.



09915-cover-quiz.jpg

Adapted from Mai Yin Tsoi

After a conversation with a colleague who was using alternative grading strategies in upper-division courses, Link started thinking about how she could change the grading in her lab courses. One sleepless night, she scoured the web for alternative grading systems and finally found specifications grading. It sounded familiar, and she remembered having caught the tail end of a talk Ring gave at the Biennial Conference on Chemical Education in 2016. At the time, she had dismissed the approach as infeasible for her large classes. But this time, she wondered if she could make it work.

“I ended up sketching out a notebook over one night,” she says. She determined how to break down the assignments into specifications. Some of her initial ideas of ways for students to demonstrate mastery turned out to be impractical because there wasn’t a good way to track them for such a large class.

She decided to try specs grading during the 2019 summer session, when the class was much smaller—only 37 students (J. Chem. Educ. 2020, DOI: 10.1021/acs.jchemed.0c00450). She worked with two experienced TAs, Kate McKnelly and William Howitz, to define the criteria that students needed to meet to achieve specific grades. The previous points-based rubrics had combined multiple items in a single rubric. They separated these into multiple rubrics that could be quickly graded on a yes-no basis.

The TAs found that the assignments were easier to grade. They also found that their discussions with students tended to focus more on student understanding than on complaints about grading.

The students had mixed reactions, Link says. They worried that some minor omission in their postlab assignments would lower their grades, but the average grades for the class were higher than in Link’s previous classes. “They are freaked out about this hypothetical situation that pretty much never happens,” Link says.

In fall 2019, McKnelly also worked with Stephen Mang, an assistant professor of teaching at UCI, to convert two of his classes to specs grading. Those classes—a writing class and an upper-level instrumental analysis lab—are significantly smaller than Link’s. “I’ve never been afraid to totally reorganize my class with 3 weeks to go before the quarter starts. I jumped in with both feet to specifications grading,” Mang says.

Mang uses a system of high pass, low pass, and unsatisfactory for individual assignments. In the writing class, he establishes criteria at the sentence, paragraph, and assignment levels, with 7 criteria for small assignments and 10 criteria for large assignments (J. Chem. Educ. 2021, DOI: 10.1021/acs.jchemed.0c00859). He grades individual criteria as met or not met, and the number of criteria that students meet determines their overall performance on an assignment. High pass is at least 6/7 or 8/10, low pass is 5/7 or 6/10, and unsatisfactory is anything lower.

“I think partial credit makes a lot of sense at the assignment level, but I don’t think it makes sense at the criteria level,” Mang says. “If we’re asking students to demonstrate mastery of a skill, that’s very much a yes or no question. If we’re saying, ‘Did you demonstrate mastery of enough skills to pass this assignment or pass this class?,’ that’s more of a fuzzy question.”

When McKnelly moved to Emory University as a lecturer, she brought specs grading with her. She’s using it in a macromolecules lab, which is the fourth course of the introductory sequence of Emory’s new Chemistry Unbound curriculum.

Even at institutions not yet using specs grading, there are signs of interest. Colleen Craig, an associate teaching professor at the University of Washington, is part of the team that teaches general chemistry. “The time is right for us to think about some of these things around grading and how we’re teaching our courses,” she says. Craig is gathering information about the approach to bring to the Chemistry Department’s undergraduate curriculum committee. She has invited Toledo to give a seminar or workshop on specs grading.

Selling points

Students have a variety of reactions to the approach, professors say.

Some students find specs grading confusing at first. They are used to taking a few high-stakes exams throughout a course in addition to more frequent low-stakes quizzes. “They don’t understand why there aren’t any exams. ‘When are we going to take an exam over this?’ is a very common question,” Wittenberg University’s Houseknecht says. The different effects of essential and general learning objectives on grades are particularly hard for students to grasp. “But once they figure it out, most of them seem to appreciate it.” Houseknecht has seen the percentage of As and Bs in his courses increase significantly since he switched from traditional to specs grading, but he’s also seen an increase in the percentage of Ds, Fs, and withdrawals. “There are students who are failing who used to be able to kind of squeeze by and pass without learning much.”

“Making sure you establish student buy-in is crucial,” McKnelly says. Midway through the semester, she did an in-class activity in which the students looked at their grade trackers to see what they still needed to do to reach their desired final grades. “I’ve noticed that every week they’ve gotten more and more comfortable” with specs grading, she says.

“I get really positive feedback from students, particularly about how they feel more in control of their learning and their grade than they did before,” Ring says.

Those two nights of cramming before exam number 2 were gone, and I think the students felt relief with that.

Mai Yin Tsoi, professor of chemistry, Georgia Gwinnett College

Students also don’t need to cram before exams as often. Instead, they have to maintain a consistent level of effort to keep on track.

“Those two nights of cramming before exam number 2 were gone, and I think the students felt relief with that,” Georgia Gwinnett’s Tsoi says. Still, the system isn’t for everyone. Tsoi remembers a student who withdrew from her class—he preferred cramming for exams because between exams, he had a couple of weeks to focus on other classes before turning back to chemistry.

And the frequent quizzes can make students think that the approach is “relentless,” especially this year, Houseknecht says. Every week they have one or two quizzes, and for the 2020–21 academic year, Wittenberg University eliminated its fall and spring breaks in response to the pandemic. “We’re all tired by the end of the semester, and I think specifications grading makes that a little harder,” Houseknecht says. “It’s not like the week after the exam you can just take a break.”

Mang’s students have fallen into distinct camps: those who like the approach and say that it’s clear what they have to do to achieve their desired grades, and those who say the system doesn’t adequately reward their effort.

Eliminating partial credit can improve interactions with students. Tsoi says it particularly changes end-of-semester conversations. “It’s never ‘Can you fix my grade’ or ‘Can you just bump me two points?’ ” she says. “It’s ‘Can I have another retake?,’ which I find fascinating. The conversation is completely different.”

Users of specs grading suspect that the approach is more equitable than conventional points-based systems. In the traditional system, some students may have difficulty interpreting what professors are looking for and may hesitate to approach faculty. Other students who are comfortable with such conversations may end up persuading professors to give them extra credit or change their grades. The transparency of specs grading “makes all that information available to all students, not just a few,” Toledo says. “And I think that’s really powerful.”

But education researchers would like to see data. Paulette Vincent-Ruz, a chemistry education researcher at the University of Michigan who focuses on issues of equity and justice, thinks specifications grading has the potential to be more equitable but stresses that equity doesn’t happen spontaneously. “You need to be really intentional about certain aspects of the design to make sure things are equitable,” she says.

One of the issues is that conventional points-based grading systems, often curved, are so prone to bias that specs grading looks good in comparison, Vincent-Ruz says. Now that more chemists are using specs grading, she says, it’s time to see whether professors can pick and choose parts of the approach they want to use and still get the benefits. She also notes that what works best may depend on the institution and its student population.

Link is collaborating with Vincent-Ruz to see if the equity claims are valid. Grades have gone up in Link’s classes since she switched to specs grading—with the most improvement in the off-sequence classes, which typically have students who have struggled in chemistry. “They are now on par with the on-sequence grades,” Link says.

Link consistently sees improved grades with specs grading. Other people might worry about grade inflation, but that’s not what bothers her. “What keeps me up at night is, ‘Was I harming the previous students?’ ”