A Quantitative Analysis of Lexical Differences Between Genders in Telephone Conversations

By Constantinos Boulis, Mari Ostendorf. 2005

53 citations recorded by the source corpus.

In this work, we provide an empirical analysis of differences in word use between genders in telephone conversations, which complements the considerable body of work in sociolinguistics concerned with gender linguistic differences. Experiments are performed on a large speech corpus of roughly 12000 conversations. We employ machine learning techniques to automatically categorize the gender of each speaker given only the transcript of his/her speech, achieving 92% accuracy. An analysis of the most characteristic words for each gender is also presented. Experiments reveal that the gender of one conversation side influences lexical use of the other side. A surprising result is that we were able to classify male-only vs. female-only conversations with almost perfect accuracy.

Subfields: Artificial Intelligence, Gender Studies, Human-Computer Interaction

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Gender Studies

In Human-Computer Interaction

Citations in the corpus

Zooming in on Gender Differences in Social Media (2016, background): “In speech, an analysis of the most frequently used words by males and females in telephone conversations was presented in (Boulis and Ostendorf, 2005), who found that swear words are more often used by males (bullshit, sucks, damn), whereas family-relation terms are more often used by females…”
Automatically Inferring Gender Associations from Language (2019, background): “Echoing Lakoff’s original claim, a popular strand of computational work focuses on differences in how women and men talk, analyzing key lexical traits (Boulis and Ostendorf, 2005; Argamon et al., 2007; Bamman et al., 2014) and predicting a person’s gender from some text they have written (Rao et al.”
Classifying latent user attributes in twitter (2010, background): “Recent state-of-the-art work in this oeuvre is by Garera and Yarowsky [7], who extend the ngram based models proposed in Boulis and Ostendorf [2] with sociolinguistic features and show applicability to a variety spoken conversational transcripts and more formal enron email.”
Modeling Latent Biographic Attributes in Conversational Genres (2009, methodology): “As our reference algorithm, we used the current state-of-the-art system developed by Boulis and Ostendorf (2005) using unigram and bigram features in a SVM framework.”
RtGender: A Corpus for Studying Differential Responses to Gender (2018, background): “…of the addressee inﬂuence linguistic behavior; existing computational work has shown, for instance, that the gender of the interlocutor inﬂuences lexical choices of a speaker in spoken and written interactions (Boulis and Ostendorf, 2005; Jurgens et al., 2017; Prabhakaran and Rambow, 2017).”
Computational Sociolinguistics: A Survey (2016, methodology): “A variety of machine learning methods have been explored, including SVMs (Boulis and Ostendorf 2005; Ciot, Sonderegger, and Ruths 2013; Corney et al. 2002; Fink, Kopecky, and Morawski 2012; Gianfortoni, Adamson, and Rosé 2011; Mukherjee and Liu 2010; Nowson and Oberlander 2006; Peersman, Daelemans, ”