How I use specifications grading in a biology seminar course
Teaching real-world science skills through primary literature
Today’s guest post is from Dr. Priscilla Erickson. Dr. Erickson is an Assistant Professor of Biology at the University of Richmond. She lives in Richmond, VA, where she enjoys birding along the James River and weekend getaways to the mountains with her husband and rescue mutt. She studied Biology at Kenyon College, has a PhD in Molecular and Cell Biology from UC Berkeley, and did post-doctoral research at the University of Virginia. Her research explores genetics and evolution in a newly invasive species of fruit fly. You can read more about her work at https://ericksonlab.net or reach her at perickso[at]richmond[dot]edu.
When I started a faculty position at The University of Richmond, I was asked to design a new upper-level biology elective with a lab. UR is a private, selective liberal arts college with Biology courses capped at 16 students, so I knew I would be able to develop personal relationships with students and use a variety of active learning strategies. All Biology students complete a three-course core curriculum and take five upper-level electives to complete the major. The elective course I developed, Genetics of Biodiversity, is a primary literature-focused seminar. In this article, I’ll describe how I’ve implemented and revised specifications grading during the three times I’ve taught this course.
I first heard of specifications grading during a workshop at UVA’s Center for Teaching Excellence during my postdoc. I read Linda Nilson’s book before I started to plan the class and learned that specifications grading generally requires students to complete some combination of more challenging work and/or a larger quantity of work to earn higher grades, with an emphasis on growth through revision. Over three semesters, I developed a scaffolded series of writing assignments that require students to interpret and analyze primary literature, paired with flexible homework assignments to assess content mastery. Students decide what letter grade they are working towards but can change plans mid-semester without penalty.
Overall course structure
Class meetings alternate between lectures and activities that prepare students for reading assignments, and discussion days with guided activities to help students interpret the data and evaluate the findings of the papers — the same skills they will be practicing in their written work. In between, students have the weekend to read and annotate the paper on the social annotation platform Perusall. Higher letter grades require an increased number and complexity of comments: for example, students working towards an A must make five comments on every paper; two comments must be directly related to data in the paper and one must be a response to another student’s comments. Before the discussion, I read all comments, respond to some, and identify areas to focus on in class. I include some flexibility in completion of these readings; students can miss 1-3 reading assignments based on their target grade or use a token to complete a makeup assignment. Students receive two tokens at the beginning of the semester and can earn additional tokens with supplementary assignments.
The lab component (3 hours per week) is a bioinformatics course-based undergraduate research experience (CURE) related to my research studying genetic variation in wild populations of an invasive insect. Each student pair does a similar lab project using essentially the same coding skills, but each pair chooses which insect populations to compare (e.g. from two different locations or two different years). Because bioinformatics is intimidating for many students, I want them to get the experience without the stress of having to recall specific programs or commands. As such, the lab activities have little influence on the overall course grade, provided they are successfully completed in class (I typically provide feedback on about 5 key questions from each week’s lab). Students do a final presentation, graded for completion, about their projects at the end of the semester. This year I used a “lightning talk” presentation format with slides that automatically advance, which kept students on topic and added a bit of excitement to what can otherwise be somewhat repetitive presentations.
Assessing course content
Part of my current assessment format was inspired by a colleagues’ introductory course with a similar structure, which you can read about here. In lieu of exams, students complete “challenge problem” sets (CPs). These are essentially take-home exams with short-answer questions that build on our paper discussions. The first time I taught the class, I used two large take-home exams that were graded pass-fail (pass meant > 65% correct), but the students found them to be anxiety-inducing, so I distributed similar questions over more assignments. Students complete a different number of CPs depending on the grade they are working towards. Each CP assignment is scored with an EMRN scale; students who earn an R or N must revise to earn credit. I grade each individual question credit/no credit, and students need to earn credit on roughly 70-75% of the questions to receive an M (a perfect score earns an E). Students work independently and sign an honor code pledge. I release a new CP each week based on that week’s material, but students can choose to complete them in any order. Students can turn in one CP on each of 10 designated weeks during the class; I do not collect CPs during weeks when another major assignment is due. The downside of this approach is that not all students are assessed on the same material, so some concepts are never assessed for some students. In the future, I may implement some required CPs covering core concepts.
A scaffolded, real-world writing project
The other main assessment in the class is a scaffolded, semester-long writing project based on “real world” science writing. I use the Transparency in Learning and Teaching format for assessment descriptions to provide detailed information about what is required to “meet specifications” and have provided links to these documents. Students choose a focal primary literature paper and complete three written assignments related to that paper. Each assignment has different expectations depending on the letter grade the student is working towards. I also provide sample student work with my feedback.
The first assignment is a detailed data summary in which students interpret 2-4 pieces of data from their paper. This year I allowed students to choose an in-person oral summary rather than a written one. These one-on-one meetings were so much more fun than grading written work! Although time-consuming, asking questions gave me a much better sense of the students’ understanding, and it helped circumvent concerns about the use of AI. The downside of this assignment is that I need to read each paper; the upside is it gets me reading papers outside my field!
The next part of the writing assignment is a critical analysis. Students identify and explain 2-4 strengths or weaknesses of their paper, incorporating at least one citation of related literature for each strength or weakness. Students find this assignment the most intimidating. They start off thinking that they, as undergraduates, have no ability to critically evaluate scientific literature but most are pleasantly surprised that they have learned the skills to be able to complete this assignment successfully.
The last piece of the project is a short research proposal based on future directions from the paper they read and critiqued. The proposal is inspired by the National Science Foundation’s Graduate Research Fellowship. I emphasize each proposal having a clearly articulated hypothesis and 1-3 “specific aims” (research directions or types of experiments) to test that hypothesis using genetic tools we discussed during the semester. Many students struggle with developing a central hypothesis for their proposal and relating each aim back to the hypothesis. Prior to writing their own proposals, students read and dissect my (far from perfect!) GRFP proposal from graduate school. We also do several in-class brainstorming activities, an outline, and an in-class peer review to get all students on track.
To earn an A in the class, students complete one final assignment: a review of an unpublished manuscript from the preprint server BioRxiv. Preprints are scientific manuscripts that are posted publicly but have not yet been through peer review. I assign preprints individually, doing my best to match student interests to paper topics and maintain a uniform length and difficulty level. Students write a review in the style of a traditional peer review to a journal editor, combining the content-related and critical thinking skills they have built all semester. I share the anonymous reviews with the authors of the papers by email, which allows students to participate in the peer review process. In the past three years, students have reviewed about 24 articles and I’ve heard back from at least 10 of the first authors, who are often impressed with the student’s work. One author wrote: “This is so helpful! We just got the reviewers' comments back, and your students’ review pointed out many parts that the reviewers missed. Please pass along a huge thank you for their thoughtful and thorough comments.” Students working towards a B or C instead review research proposals written by other students in the class, a substantially smaller assignment that uses similar skills.
Below is a chart that I use in my syllabus to summarize the requirements for each grade level. Students need to successfully complete all requirements to earn a grade in a given range.
Revisions
Revision is a key part of the feedback cycle in specifications grading. I provide detailed feedback on all written work, both with in-text comments and summary statements on every assignment. For the challenge problems, students can turn in one revised challenge problem each week after the first week. They can revise the same problem multiple times, and they can use a token to turn in an extra revision in any given week.
For the summary, critique, and proposal, students have one automatic revision. Students can use a token to submit additional revision attempts, but this is rare; nearly all students meet specifications with a single revision attempt. The final preprint review does not allow time for a revision.
Metacognition and grade determination
Students complete self-reflections at the beginning, middle, and end of the semester. The first one allows me to get to know them and their goals; the second two are for them to assess their progress on the learning outcomes of the class, their contributions to the classroom, and their work habits.
Students can have a say in their semester grade through these self-reflections. Their completed assignments put them in a letter category based on the table (i.e. B-range), and students can advocate for a fractional grade within that category. I typically give the grade they suggest. I also typically use plusses for “close misses”; for example, if a student has all the requirements for an A but is missing one challenge problem at the end of the semester or had a major issue with one aim of their grant proposal, they would earn a B+.
Closing thoughts
After teaching this course three times, I am generally happy with specifications grading, though I do spend substantial time commenting on student work. Feedback from students generally indicates that although the class is demanding, it is less stressful than an exam-graded class. Most students genuinely appreciate the opportunity to practice real-world science skills and to revise their work. Some who have experienced this form of grading before say they are now disappointed when they take classes that don’t allow revision. The response from my colleagues has likewise been supportive. I recently completed my third-year review and the departmental letter pointed out the success of the system as evidenced by student evaluations. Although grading multiple submissions of each assignment can be a lot to manage, I enjoy working with students individually and appreciate not having to agonize over whether an essay earned an 88% or a 90%. I try to create clearly defined expectations and just have to decide if students did or didn’t meet each one.
If you're interested in doing something like this yourself, here are some key pieces of advice:
Social annotation platforms are a great way for students to engage with primary literature and each other before class. I require more comments (and specific types of comments) for students working towards higher grades.
If you are using writing assignments, consider ways to make the assignment more challenging or more complex for students wishing to earn higher grades (i.e. more arguments, more experiments, and/or more citations)
Consider using a series of self-paced assessments; successful completion of more assessments corresponds to higher grades in the course. Keep deadlines evenly spaced throughout the semester (with opportunities for flexibility) and avoid letting work pile up at the end. I use Blackboard to track student progress using many of the strategies described here. I also remind students of upcoming deadlines at the beginning of every class.
A final assignment for students working towards an A can be a meaningful experience. A smaller substitute assignment for students working towards lower grades can balance the workload.
Incorporate metacognition when possible to remind students of what they are doing and why.
Acknowledgements: Thank you to Dr. Melinda Yang for many thoughtful conversations about the design of this course and to Dr. Kylie Korsnack for helpful feedback on this post. The BioRxiv review assignment was inspired by this article. The other writing assignments were inspired by a course I took with Dr. Karen Hicks as an undergraduate.
 | A guest post by
|