Examining Gender Differences in Language: A Computational Analysis of Emotion in the Dutch Veteran Institute Oral History Archive

Abstract

Gender differences in post-traumatic stress symptoms suggest that expressing trauma is influenced by gendered socialization, as gender stereotypes influence the way people express their emotions. In this research, emotional language in 17 interviews of the Dutch Veteran Institute oral history archive was analyzed using applications of artificial intelligence. In these interviews, conflicting and traumatic events were recalled. The interviews were transcribed using Automatic Speech Recognition technology, and analyzed using a keyword-based approach using the Dutch LIWC2007 and LiLaH emotion lexicon. The objective of this research was to examine gender differences in the use of emotional language, and to see whether these differences fell in line with gender norms for emotional expression. It was found that women used more words related to joy, friends and humans, whereas men used more swear words and words describing positions in space. The results show that emotional gender stereotypes are slightly present in the language of this dataset.