Self-Evaluation: Turning Evaluations into Learning Opportunities

In Self-Graded Homework Helps Students Learn and Promotes Accurate Metacognition, 3 Psychology professors from the University of Delaware examined the use of self-evaluation with their students. Each student had to grade themselves on half of their homework assignments (after first discussing their answers with peers in groups and having access to the written answer keys). For the evaluations that followed these self-graded homework tasks (and covering the same topics), students’ grades were higher than in evaluations following traditionally graded homework.  

The logical explanation is that self-evaluation encourages students to reflect on their errors (and their strengths) during homework, leading to better learning outcomes.   

Do students really read our feedback?  

I’ve always been a diligent student. From elementary school to university, I religiously reviewed the corrected assignments my teachers returned, carefully reading all their comments.  

In contrast, my 10-year-old son surprises me every week with his indifference to his teacher’s feedback. He shows me his weekly evaluations and is aware of the grades he has received. He’s sometimes disappointed with the results, but when I ask him what went wrong, he often says he doesn’t know or gives an answer that contradicts his teacher’s comments. I ask him if he’s read her feedback, and he says “no.” I then read it with him … only to repeat the same scenario the following week!  

I’ve come to realize that, unfortunately, many of my CEGEP students behave more like my son than how I used to. 

Ideally, evaluations should serve as learning opportunities, yet if students are unable to identify what they can do and what they still need to learn, we fall short … 

Study conducted at the University of Delaware  

A research team from the University of Delaware (Grace Handley, Anton Lebed, and Beth Morling) conducted a study in a course that featured 4 exams, each preceded by a homework assignment. For half of the groups, the homework leading up to the 1^st exam was self-graded, while for the other half, a teaching assistant marked the homework as usual. For the assignment before the 2^nd exam, the conditions were reversed: the 1^st half received traditional marking, and the 2^nd half graded themselves. This alternating pattern continued for the 3^rd and 4^th exams. 

After submitting a homework assignment that was not self-graded, students discussed selected questions in small groups. The teaching assistant clarified challenging questions and encouraged students to review their answers using the answer key posted on the course’s learning management system. 

For the self-graded homework, students were asked to evaluate their work by comparing it to an exemplary answer displayed on a screen at the front of the classroom and identifying both conceptual and structural differences. The assistant teacher explained the correct answers and guided the students on how to use the grading scale for each question. 

For each question discussed in class, students gave themselves a score from 1 to 3:  

• 3: attempted and fully correct 
• 2: attempted but with some mistakes  
• 1: attempted but entirely incorrect 
• 0: not attempted 

Not every question on each assignment was necessarily discussed in class and self-graded.  

The teaching assistant went over each self-assigned score and adjusted it where necessary (focusing more on completion than accuracy). The assistant then gave an overall grade for the entire assignment. In the 1st semester of the study, the grading system was out of 100, but students found this stressful. For the 2^nd and final semester, a 0-to-3 scale was used instead to better reflect the fact that each homework grade represented only a small fraction of the final course grade. (Homework grades were included in the final grade, though I do not know the exact weighting.) 

Results 

Impact on exam grades  

The exams included both multiple-choice and open-ended questions.  

The self-evaluation did not significantly impact the grades on multiple-choice questions. This had been predicted by Handley, Lebed and Morling, since the homework assignments consisted of open-ended questions rather than multiple-choice questions.  

However, students who had self-graded their homework performed significantly better on the corresponding open-ended exam questions, but only on the first 2 exams. By the 3^rd and 4^th exams, the difference was no longer significant. This can be explained by the fact that by then, all students had already experienced self-evaluation, and some continued to self-evaluate even when it was not required.  

Impact on how accurately students could predict their exam grades 

Handley, Lebed, and Morling also observed that students were better at predicting their exam grades when they had self-graded the previous homework. Interestingly, this metacognitive effect was observed across the 4 exams, unlike the impact on their actual grades. 

A transferable approach? 

In this study, self-evaluation was used in a summative evaluation context: the grades students assigned themselves contributed to their final course grade (after being reviewed by a teaching assistant). In my own teaching context, such an approach might be difficult, but I could use self-evaluation in a formative context. Even if some students might not take it seriously (or might skip it altogether), I’m convinced that several would benefit from the experience. It’s worth trying! 

If you use self-evaluation in your teaching or have other strategies to turn evaluations into real learning opportunities, we’d love to hear from you! Share your experiences in the comments below!