Mitigating Teacher Biases in Assessment and Grading

• In Event: Saturday Roundtable Session 9:45 am
  In Roundtable Session: The Teacher Side of Assessment: Exploring Biases, Beliefs, and Attitudes (Table 11)

Abstract

Among the many skills teachers are expected to possess, providing high-quality feedback to students and grading their work are critical yet complex tasks (Parr & Timperley, 2010). Empirical evidence shows substantial variability among teachers in grade assignment, regardless of the work's quality, grading approach, or specialization area. Further, implicit teacher biases and stereotypes concerning gender, race/ethnicity, and migrant background significantly influence student outcomes and grading, with evaluation bias impacted by the characteristics of the evaluation tool (Quinn, 2020). Despite progress in the field, there is a scarcity of studies explaining factors contributing to bias in grading and feedback among pre-service teachers. This study aims to address this gap, guided by the following research questions:
1.Do teacher candidates differentially grade male and female students’ essays? Do these potential differences persist when they employ a different grading method (holistic vs. analytic)?
2. Do students’ and teachers’ gender alignment predict the quality of student feedback comments?
Method
This randomized experimental study was conducted in a sample of 174 college students (56% female) enrolled in a teacher education program in Spain. The participants were presented with a standard essay and were instructed to provide feedback and grade. The student’s name (Valeria vs Alejandro) was manipulated and used to sign the gender of the essay author. The grading method (holistic vs. analytic) was also manipulated. Hence, the participants were randomly assigned to one of the four conditions (girl/analytic, girl/holistic, boy/analytic, and boy/holistic).
Results
The ANOVA showed a statistically significant main effect of teachers’ gender (F-test= 3,888, p = .050) and an interaction effect between the essay author's gender and the grading approach (F-test= 6.700, p= .010) (Table 2). Pairwise comparisons using Benjamini-Hochberg correction revealed that female teachers rated the girl’s essay (Mean = 5.42, S.D= 1.22) higher than the boy’s (Mean= 4.43, S.D = 1.59) when using an analytic approach. In contrast, under the holistic method, women participants rated the boy’s essay (Mean = 5.12, S.D = 1.25) higher than the girl’s (Mean = 4.28, SD =1.85) (Table 3). Furthermore, women participants rated the girl’s essay higher under the analytic grading approach than the holistic method. We found no significant statistical differences between conditions for male participants.
A total of 816 feedback comments were coded by two researchers. Significant differences in proportions of feedback depending on teacher/student gender combinations were observed (Table 4). For example, participants with the same gender as the student provided a higher proportion of task and future-oriented messages, and a lower ratio of praise or process subcategories than teachers of the different gender with the student. Additionally, female teachers were more consistent than male teachers in the proportion of messages provided to both girls and boys except for praise, where the percentage of comments in this category was higher for boys.
Significance
Examining teacher biases and ways to mitigate them can help shed light on potential inequities and inform strategies to ensure fair and unbiased assessment practices for all students.

Authors

• Anastasiya A. Lipnevich, City University of New York

• Carolina Lopera-Oquendo, City University of New York

• Juan Fraile, Fundación Universidad Francisco de Vitoria

• Patricia Ruiz Bravo, Universidad francisco de victoria