Why I give (almost) all of my assessments back to my students
How important is it that students get their work back in alternative grading?
Sharona Krinsky teaches Mathematics at California State University, Los Angeles. She uses Alternative Grading in a wide variety of classes and program structures, including as the coordinator of a general education statistics course with over 1,500 students per year in over 70 sections and with 15 other instructors. She is a founding organizer of The Grading Conference, now in its fourth year. Sharona has her master’s degree in Mathematics from The Ohio State University. When not teaching math or working to get Alternative Grading adopted, she is the Executive Producer of an educational theater company serving the South Bay area of Los Angeles (visit www.EncoreSouthBay.org to see what is coming next!).
As has been written about extensively on this blog, Helpful Feedback is the second pillar of the Alternative Grading Framework (see The care and feeding of Helpful Feedback, Planning for grading for growth: Fleshing out the feedback). When discussing feedback, giving it and helping students use it, one question I am often asked when working with faculty is, “Do I have to return quizzes/exams/assessments to students?” This came up recently in a Twitter conversation and was also a topic of conversation on the Alternative Grading Slack channel. You might be asking, “Are some people thinking that you shouldn’t hand back work?” Yes, there are some people who think that. Not only are they thinking that, they are acting on it. This post is a direct result of those recent conversations. In this post, I will lay out my perspective on why returning most, if not all, assessments to students is a critical component of the effectiveness of the Helpful Feedback pillar.
First, let’s set some expectations. I am not a scholar of the science of learning. I have greatly enjoyed learning about learning, and have a whole series of books and articles I can recommend for those interested in understanding more about the how and the why of human learning. Second, although I am working on multiple research projects to quantify the impact of alternative grading on student success, the recommendations I am about to share are primarily from my personal experience working for over 30 years as an educator in higher education and as a parent to two college students.
What Is An Assessment
For the purposes of this post, I will define an assessment as any task, interaction, or other piece of evidence that I use to check for a student’s current level of ability to meet my expectations for a skill or task. In my classes, assessments usually consist of a Canvas quiz or a longer task in the form of a portfolio style assignment or project such as a paper, presentation or video. I distinguish an assessment from other types of student work in that I’m seeking to see if a student has met my expectations of skill versus giving students the opportunity to develop that skill. To be more concrete, a “homework” that is not used for assessment is an assignment where students usually have access to answers or get credit for completing the assignment without using a “correctness” criterion. An “assessment” is an assignment where I am specifically looking at the material provided for correct work, complete work, and/or work that overall meets the expectations that have been given ahead of time to the student.
What I Do With Assessments
Depending on which class I am teaching, I am either using Standards-Based Grading or Specifications Grading (or a blend of the two). For more on the specifics of what these grading systems entail, see David Clark’s post on this blog - An Alternative Grading Glossary. In either case, students submit work (either handwritten, typed, recorded via video, etc.) and I evaluate it. An example from linear algebra might be a quiz consisting of the following problem:
Students will usually submit scans of their handwritten work, uploaded into Canvas. I will review their work, provide detailed feedback as to the completeness of their explanations, comments and questions regarding any mathematical errors. Then, depending on the quality of the work, students will receive one of three rubric marks that represent the levels of “Complete”, “Revise” or “Retake”. The “Complete” level means that this specific assessment has completely met the expectations of the quality of work I am looking for. Students need to get two assessments “Complete” in order to achieve a “Complete” on the associated standard. Once the student has completed the associated standard, they do not need to submit any additional assessments on it.
Students who receive a “Revise” mark need to revise the work that they submitted on this specific assessment. There are guiding questions or detailed feedback on where they went astray and what additional work I am looking for for this specific assessment. Students who receive a “Retake” mark will need to try a completely new assessment on this same standard, as their current work is not of a sufficient quality. Usually, students who receive a “Retake” have significant conceptual errors that I believe need additional learning before they will be able to complete the work to a sufficient quality. Students still receive detailed feedback on the work submitted including guiding questions or specific indications of incorrect or insufficient work.
Why Give Back Student Work
This is the point at which, when I’m speaking about my grading architecture, I begin to get the questions from instructors about handing back student work. Do I actually give back the assessment questions? But what about having to do all that work to write new assessment questions all the time? Isn’t it enough that I provide feedback to students about what they did wrong? They don’t really need the original questions and their own answers back, do they?
Here are my answers. Yes, I actually give back the full assessment questions. AND the student work. AND detailed feedback about their errors. Yes, I do have to write new assessment questions (more on that later). No, it’s not enough to provide feedback; they really do need the original questions and their own answers. Why? First, consider some other situations where feedback is common.
If you have ever been around live theater, the most effective time to give “notes” (feedback from the creative team to the actors) is either immediately after a scene, when it is fresh in everyone’s memory, or upon playback of a recording of a scene. Trying to internalize a bunch of feedback after a three to four hour rehearsal is much more difficult. Oftentimes, actors make choices “in the moment” and may or may not remember which choice was made in what scene.
If you have ever submitted a paper for publication or a grant application, what would it be like to get reviewer comments but without access to the paper that was submitted? This has happened to me when I submit grants through online portals and for some reason don’t keep a copy of the most up-to-date version.
You can see how difficult it is to understand feedback on material that you already know well when you don’t have the original work in front of you, or when feedback is not given immediately. How much harder do you think it is for students to do the same task, but on material that they are just learning?
Here is what I have come to realize about students.
College is overwhelming. The amount of data students have to process on a daily basis is unbelievable. I encourage you to sit down with a student and have them show you THEIR view of the LMS system, with its four (or more) classes, every one of them structured differently, with a to-do list a mile long. Then layer on top of that the rest of life, including learning to be an adult, navigating social situations, working, family.
Assessments are stressful. Regardless of how much we try to structure our classes to be less stressful, students are under tremendous stress to succeed. Everything from financial aid (that pays for classes but often also pays for housing and food) to parental expectations, job requirements and more centers around grades and passing classes. As a result, students are under stress while taking assessments. Significant literature has shown that stress impacts the ability of the brain to record and retrieve memory. Some examples can be found here, here and here. Therefore, students often cannot recall the material from the assessment.
Students are, most of the time, very inexperienced with learning from feedback. Most traditional grading systems do not provide the opportunity to improve a grade through learning from feedback. Additionally, students have not been specifically taught to learn from feedback.
Looking at these three points in combination has led me to believe that students must have the original assessment, their own work, AND my feedback to work with when attempting to learn from their mistakes. They have absolutely no recollection of the problems or their own work from the time of the assessment. Since they don’t always know how to use the feedback to correct their mistakes or misunderstandings immediately, they need TIME to sit with the feedback, reflect on it, and attempt to comprehend it. When any of these pieces are missing, students are largely unable to deconstruct their own thinking at the time of taking the assessment, compare it to the feedback given, and repair their understanding. Even with all of these pieces students often need additional guidance, usually given in the form of a conversation with leading questions, to walk through the comparison between their work, the “correct” answer (if available), and the feedback.
This point was recently driven home to me even more by working with one of my sons. He is taking Linear Algebra and Differential Equations this semester (one of the classes I also teach) but his class uses a traditional grading structure. He really wants to understand what he is learning. Recently, as we were working together to prepare for his first midterm, I asked him if he had gone over his first quiz (that quite frankly, he did not do as well on as he had hoped). He then informed me that although he got the score back on the quiz, the instructor had not returned the quiz, nor gone over it in class. So he had no clue what he had done right or wrong.
How is any student supposed to learn from their mistakes if they do not have access to those mistakes to learn from? How is any human being supposed to learn from their mistakes if they do not know what those mistakes are?
But What About CHEATING?
This is probably the biggest point I hear from instructors as to why they do not return student work. It comes in various forms:
“I do not want my exams to be in the files of the fraternity or sorority.”
“It is unrealistic, given my workload, for me to write new assessments all the time.”
“Writing good, authentic assessments in my field is so hard that there are an extremely limited number of good problems.”
In all honesty, I have a tremendous level of sympathy for the reasons given above (especially about workload). However, they all boil down to one thing: I think my students are going to cheat more if I provide my exams/quizzes/assessments back to them. So we have to ask, are they really? Or are the students who are willing to cheat going to, and the ones who are not willing to won’t? I don’t have a clear answer to that, exceptto refer back to this post: Does alternative grading make cheating more likely? In his post, Robert explores the research on academic motivation. After reviewing the research, Robert concludes that alternative grading reduces the incentives to cheat when compared to traditional grading. As a follow-on to his post, I posit the following: giving students back their work and giving opportunities to learn from the feedback will help build students’ self-efficacy which can turn into less cheating, regardless of the overall grading structure. In other words, by giving students their own work along with helpful feedback and teaching them how to learn from it, we increase their concept of their own ability to learn, which in turn leads to their personal investment in their own learning.
On the other hand, what is the alternative? There is an essentially endless rabbit hole that you can dive into to try to stop students from having access to previous work. From timed in-class assessments or proctoring software, to writing twenty-five versions of a quiz, to putting “gotcha” items in each new version of the quiz that are irrelevant but are checking to see if they are doing their own work. However, three things come up in this situation. First, if you go down the road of policing access, there is no winning. You will always be fighting to get ahead of it. Second, each step down this road shows less trust for students, which means they will have correspondingly less interest in earning your trust. This is a new feedback loop, a loop of distrust, which creates an atmosphere of antagonism. It’s us versus the students, and therefore cheating is more OK. Third, it’s a TON of work. In fact, it’s more work (and more disheartening work) than just writing a new assessment.
I’m not being naive about this: I do have students cheat. And when I catch them, my response is simple - they don’t get credit for the work and I don’t give feedback. I communicate with my students that the most valuable thing they get from the assessments is my feedback. When it is taken away, that is the single most significant consequence (in addition to not getting credit for the work). Once they begin to learn the value of the feedback, that alone reduces the temptation to cheat.
What About the Difficulty of Writing New Assessments?
This is definitely the part of this discussion for which I have the most sympathy. For this reason, I am absolutely thrilled that more and more resources are coming online to help with writing additional assessment questions. I personally use CheckIt, an OER tool that gives tremendous capabilities to instructors to author questions with a ton of variations. Using CheckIt, I am able to generate literally hundreds of versions of a question, which allows me to let over a thousand students take an asynchronous Canvas quiz and minimize the impact of things like ChatGPT and Chegg. Canvas itself has some ability to do algorithmic versioning, as do most of the online assessment tools from the major textbook publishers. To explore this feature, see some of the Canvas help documentation such as this page about creating Formula quiz questions.
For a less computation-oriented class like History of Mathematics, the assessments become more about the open-endedness of the prompt and the design and structure of the assessment. The more choices that I have students make (and tell me what those choices are) the more likely they are to provide authentic, interesting work and I don’t have to do the work of versioning. For example, a learning outcome in History of Math might be “I can describe ways in which gender and race have influenced who identifies as a “mathematician” both historically and in modern time.” Then, the specific assignment asks students to use historical information and people to explore this content. Students have tremendous leeway in the eras and geographies that they might consider, from Ancient Greece to 17 or 18th century China or even to the 21st century. Due to the open-ended nature of the assignment, I don’t have to make a lot of versions, students make their own choices based on their personal interests. This lets me use their own work as well as the revision of that work as an authentic assessment.
At the end of the day, what is the most important reason we teach? For me, it is to help students learn. I have to give grades, but I want students to learn. Even more, I want students to learn how to learn. If I don’t give them back their work, I cut off their ability to do both. So regardless of the grading system you use, I implore you - please give students their work back, with feedback.
Thanks for reading Grading for Growth! Subscribe for free to receive new posts in your inbox every Monday.
In fact, this post is what happens when you privately complain to your friend, who then says, “Sounds like a great blog post. You’re going to write it, right?”
Image taken from a public problem located at https://checkit.clontz.org/
This usually happens when I go over the character count and have to trim on the fly.
If you want to get really disgusted about proctoring software take a look at some of the reporting on students’ experiences with proctoring software such as this one https://www.insider.com/viral-tiktok-student-fails-exam-after-ai-software-flags-cheating-2020-10
Full disclosure, the one piece of work that I don’t return is final exams. Students are done learning (at least within the course), so I do reduce my workload a little by not returning finals :)
This is a very very well-argued post. I agree that constantly generating additional assessments is a hassle AND that it's what we need to do to maximize our students' opportunities to learn. CheckIt sounds like a fantastic tool, but I think it's math-specific. For biology I've developed something called Test Question Templates (TQTs), which are not slick and automated but make it somewhat easier for faculty to spin off many variations of a problem (as summarized at my 2022 Grading Conference talk -- thanks for organizing!).
I agree with giving assessments back for feedback. I also do not give final exams back for the same reason that no student cares about the feedback except to contest the grade if they are failing.
By not giving feedback on midterms and quizzes, we are missing out on why assessments are critical in learning.
There are about 30 years of AP calculus exams out there in the web and I do not see more people aceing the AP exam. Once there are 30 years of exams floating around, more power to the student if they do them all and hence do better on the exam. It is even more work than the HW assigned to the students.
What is bothersome is when current exams are floating around which have not yet been taken by others such as in mastery grading. That surely needs attention.