Last Names Linked to Grading Bias: Students with Surnames at the End of the Alphabet Receive Lower Marks
A groundbreaking study conducted by the University of Michigan has exposed an alarming trend in grading practices: students with last names closer to the end of the alphabet receive lower grades compared to those with earlier surnames. This bias emerges when grades are assigned alphabetically, a common default setting in learning management systems like Canvas.
After analyzing 30 million grading records, researchers discovered a consistent pattern. Students with last names falling between A and E scored 0.3 percentage points higher when graded randomly. However, those with surnames near the end of the alphabet faced a 0.3 percentage point deduction, resulting in a significant 0.6-point gap.
While this disparity may appear minimal, the study highlights its potential impact on grade point averages and subsequent career paths. Notably, for a small group of graders who graded assignments in reverse alphabetical order (Z to A), students with last names at the end of the alphabet exhibited higher marks due to their position at the top of the list.
Researchers attribute this bias to unconscious cognitive fatigue among graders. As grading assignments progress, graders may experience decreased attention and cognitive abilities, leading them to favor students with earlier surnames.
To address this issue, researchers recommend changing the default alphabetical setting in grading systems to randomize student order. Additionally, hiring more graders to distribute the workload and training graders to recognize and overcome potential biases can help mitigate this grading disparity.
