Code and Decode: A Computational Model of Colonial Language Bias in High School World History Textbooks (Poster 23)

• In Event: Friday Poster Session 3:05 pm
  In Workshop: AERA Youth Teams in Education Research Poster Session

Abstract

This study endeavors to use computational modeling to dissect the biases inherent in world history textbooks, particularly concerning representations of colonialism, and their impact on student understanding. By employing computational modeling techniques, the study aims to quantify and analyze these biases at scale. The algorithmic approach enables the examination of vast amounts of textual data from multiple sources, facilitating the identification of sentiments, biases, and language clusters. The algorithm developed for this study parses text sentence-by-sentence, categorizing words and phrases based on predefined dictionaries. Postcolonial theory, which views textbooks as institutional discourse, informs the choice of textbooks as a critical part of student engagement with world history content. Preliminary findings revealed a prevalence of pro-imperialist language in sections pertaining to British colonization of India, underscoring the need for more nuanced analytical tools. While the initial stage of the study served as a proof of concept, it also highlighted the necessity of refining the algorithm to capture the complexity of colonialist discourse accurately. To enhance the algorithm's efficacy, the study expanded to a collaborative effort to create more comprehensive dictionaries, leveraging the collective insights of students gathered through Google form data. Ultimately, this study aims to challenge the process by which textbooks and other educational resources are written by scholars and used by teachers in high school classrooms across the country.

Authors

• Amy Pozza, Saint Mary's Hall

• Helen Trottmann, Youth Teams in Education Research

• Sohan Bhakta, Youth Teams in Education Research