Investigating Racial Bias and Attribution Error in Grading Student Performance

Keywords

Grading, Standards-based Grading, Competency-based Grading, Cognitive Bias, Racial Bias, Attribution Error, Heuristics

Abstract

Standards-based grading (SBG) is a relatively recent alternative model of grading designed to increase the reliability of grading while reducing the influence of teacher biases that can distort student evaluations. This study investigated whether SBG mitigates race‑related bias and attribution errors through a mixed‑methods 3×2 factorial vignette experiment with in‑service teachers. A total of 219 teachers scored the same student essay using a common SBG rubric and provided a 1-4 proficiency score, a binary evaluation concerning success criteria and written feedback. “Learning profiles”– a description of the student’s current and past abilities – varied by tone (positive, negative, none: name-only) and implied race (e.g., Jake/Jamal); vignettes were randomized in a 3 (tone) × 2 (race) design. Proficiency scores and success criteria evaluation were analyzed using ordered logistic regression, and written feedback was coded according to Weiner’s attribution dimensions to reveal how teachers’ evaluative reasoning operates under SBG.

First, learning profile tone showed a relationship with rubric scores (positive > name‑only; negative < name‑only), whereas the profile and implied race interaction showed no significant effect. Regarding feedback, across conditions, the most common attribution was internal–unstable (malleable ability/effort). Secondary patterns indicated that Jake profiles elicited more internal–stable language (you have the ability), whereas Jamal profiles elicited more external–stable references to standards/rigor (writing must meet a rigorous standard of quality).

Findings of this study include, although SBG’s structure aims to objectify grading, it remains susceptible to certain biases. Importantly, we interpret the null race effects on scores as no detected differences within this design rather than evidence that SBG eliminates racial bias. Feedback, however, showed patterned attributions across name and profile tone, indicating opportunities for attribution‑aware feedback training. Majority of feedback aligned with SBG’s growth orientation (internal-unstable) yet there were higher than expected attributions aligned with Wiener’s internal-stable (fixed‑ability) and structural (external-stable) attributions persist, underscoring the need to refine scoring and feedback practices to promote objectivity in student assessment.

This paper contributes to current discussions on grading reform, stressing the importance of refining grading methods to ensure fairness and objectivity in evaluating students.

Recommended Citation

Reibel, Tony and Roegman, Rachel (2026) "Investigating Racial Bias and Attribution Error in Grading Student Performance," Mid-Western Educational Researcher: Vol. 38: Iss. 1, Article 9.
DOI: https://doi.org/10.25035/mwer.38.01.09
Available at: https://scholarworks.bgsu.edu/mwer/vol38/iss1/9