Bias in the Writing of Prose and Its Appraisal

Evidence from a variety of sources suggests that systematic differences can be found in the ratings given to student essays as a function not only of the student’s skills but also of aspects of both the student’s background and the background of the rater. Additionally, the nature of the prompt which provided the central theme of the essay might bias the outcome of the ratings of that essay. A study of ratings of fifth and sixth graders who wrote paragraph-long essays in response to two topics presented either in written or pictorial form is presented. Students were classified as Hispanic-surnamed or non-Hispanic-surnamed; two teachers, trained as raters using an objectively-based essay scoring scheme, represented an Hispanic cultural background and two a non-Hispanic background. Results from a blind rating of 100 complete essays show that several of the rating subscales were significantly influenced by an interaction between student ethnicity and rater ethnicity, and several subscales by rater ethnicity alone. Student ethnicity alone was not a significant main effect on any subscale. Prompt modality is significant for one subsca1e, and interacts with rater ethnicity on one other. The findings are interpreted as a direct indication of biased assessment.