Home > Conventions > Program Archive > APS Virtual Poster Showcase > Presentation
APS

APS Virtual Poster Showcase · 2020

Assessing the Prevalence and Strength of Gender Associations in Children’s and Adult’s Language

Virtual · June 2020

Poster Sessions · Social

  • Alex Youn
    Harvard University
  • Elizabeth Wu
    Harvard College
  • Tessa Charlesworth
    Harvard University
  • Mahzarin Banaji
    Harvard University

Abstract

Using natural language processing (word embeddings), this study analyzes child-produced, child-directed, and adult-produced text to quantitatively compare the prevalence and strength of nine gender associations (e.g., male-career/female-home). Results indicated that gender associations from adult books displayed the strongest overall effects, while parent’s speech displayed the weakest overall effects.

Statistics and Methodology