Reading between the Lines: Revealing and Resisting the ‘Hidden’ Gender Bias in Math Story Problems

Abstract

Educational testing has a long and complicated history in the United States. Currently, state- and district-level assessment systems shape the educational process in a number of ways, from curricular design to teacher evaluation to funding allocation (Au, 2007, 2011; Condie, Lefgren, & Sims, 2014; Harris, 2012; Madaus & Clarke, 2001; Medina & Neill, 1990; Sass, Semykina, & Harris, 2014). This work explores tests as both a measure of and a mechanism in the maintenance and spread of inequity. In the first study, ‘gold-standard’ math story problems are qualitatively coded by three reviewers for the presence of gender bias and stereotypes to reveal patterns in representation of gender. In the second study, a text classifier is developed to parse math story problems by gender and activity. The final piece presents math teacher reflections on making math and math assessment more equitable for students and offers a framework and resources for creating, modifying, and discussing problematic test questions with colleagues and students. The three papers together begin a discussion regarding the goals and impacts of culturally embedded assessment and challenge many of the assumptions made in past and current research about gender differences in math, what makes assessment equitable, and how we can move forward towards academic justice.