Gender identity and lexical variation in social media

By David Bamman, Jacob Eisenstein, Tyler Schnoebelen. 2012

446 citations recorded by the source corpus.

We present a study of the relationship between gender, linguistic style, and social networks, using a novel corpus of 14,000 Twitter users. Prior quantitative work on gender often treats this social variable as a female/male binary; we argue for a more nuanced approach. By clustering Twitter users, we find a natural decomposition of the dataset into various styles and topical interests. Many clusters have strong gender orientations, but their use of linguistic resources sometimes directly conflicts with the population-level language statistics. We view these clusters as a more accurate reflection of the multifaceted nature of gendered language styles. Previous corpus-based work has also had little to say about individuals whose linguistic styles defy population-level gender patterns. To identify such individuals, we train a statistical classifier, and measure the classifier confidence for each individual in the dataset. Examining individuals whose language does not match the classifier's model for their gender, we find that they have social networks that include significantly fewer same-gender social connections and that, in general, social network homophily is correlated with the use of same-gender language markers. Pairing computational methods and social theory thus offers a new perspective on how gender emerges as individuals position themselves relative to audiences, topics, and mainstream gender norms.

Subfields: Artificial Intelligence, Experimental and Cognitive Psychology, Social Psychology

View Paper: https://www.semanticscholar.org/paper/5eec359ca2903a20700019761a6e2f843d4bec6b

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 Experimental and Cognitive Psychology

In Social Psychology

Citations in the corpus

Obfuscating Gender in Social Media Writing (2016, background): “Even though this binary may not be an accurate reﬂection of the gender performance of users on social media (Bamman et al., 2014; Nguyen et al., 2014), we operate under the presumption that most demographic classiﬁers also use two genders.”
Gender Bias in Machine Translation (2021, unknown): “Otherwise, if we assume that the only relevant categories are ‘‘male’’ and ‘‘female’’, our models will inevitably fulfill such a reductionist expectation (Bamman et al., 2014).”
#thyghgapp: Instagram Content Moderation and Lexical Variation in Pro-Eating Disorder Communities (2016, background): “[4] extended these investigations of lexical variations in Twitter to gender identity.”
Story Cloze Task: UW NLP System (2017, unknown): “(Argamon et al., 2003; Schler et al., 2006; Bamman et al., 2014), and native language (Koppel et al.”
Social and Emotional Correlates of Capitalization on Twitter (2018, background): “Gender lines divide language use in speech (Eckert and McConnell-Ginet, 2003); in writing (Koppel et al., 2002); and on social media (Koppel et al., 2006; Bamman et al., 2014).”
Gender and Power: How Gender and Gender Environment Affect Manifestations of Power (2014, unknown): “This notion of gender environment is supported by an idea in recent Twitterbased sociolinguistic research on gender identity and lexical variation (Bamman et al., 2014).”
It’s going to be okay: Measuring Access to Support in Online Communities (2018, unknown): “Gender can also manifest through more subtle, stylistic cues (e.g., Schnoebelen, 2012; Flekova and Gurevych, 2013; Bamman et al., 2014; Volkova et al., 2015; Garimella and Mihalcea, 2016; Carpenter et al., 2016).”
A Survey on Gender Bias in Natural Language Processing (2021, background): “On the other hand, interpretation bias is a phenomenon of researchers assuming that gender is a relevant variable which ultimately leads to analyses that are incapable of revealing violations of this assumption [Bamman et al. 2014; Koolen and van Cranenburgh 2017].”
Using matched samples to estimate the effects of exercise on mental health from twitter (2015, background): “Given the observed lexical variation between genders on social media (Bamman, Eisenstein, and Schnoebelen 2014), this will confound hypothesis testing , since many factors other than exercise are contributing to the observed difference in mood.”
Detecting Cross-Geographic Biases in Toxicity Modeling on Social Media (2021, methodology): “We also conducted experiments using (Monroe et al., 2008) method in a one-vs-rest manner (described above), as well as using (Bamman et al., 2014) method for both steps; but they provided qualitatively worse results than the above approach.”
Unsupervised Discovery of Gendered Language through Latent-Variable Modeling (2019, methodology): “…how word choice in tweets is inﬂuenced by the gender of the tweeter; Schoﬁeld and Mehr (2016) modeled gender in ﬁlm dialog; and, in the realm of social media analysis, Bamman et al. (2014) discussed stylistic choices that enable classiﬁers to distinguish between tweets written by men versus women.”
Language-independent Gender Prediction on Twitter (2017, background): “Gender prediction is a well-established task in author proﬁling, useful for a series of downstream analyses (Schler et al., 2006; Schwartz et al., 2013; Bamman et al., 2014) as well as predictive model improvements (Hovy, 2015).”
Computational Sociolinguistics: A Survey (2016, methodology): “Females tend to use more pronouns [6, 7, 8, 17, 194, 231, 233] (in particular first person singular [186, 194, 233]).”
Challenges of studying and processing dialects in social media (2015, background): “…that may be relevant for other studies of dialect in social media, is almost completely irrelevant for our study: It is often important to control for topic in dialectal and sociolinguistic studies (Bamman et al., 2014), e.g., when studying the lexical preferences of speakers of urban ethnolects.”
Key Female Characters in Film Have More to Talk About Besides Men: Automating the Bechdel Test (2015, background): “There has been much work in the computational sciences community on studying gender differences in the way language is used by men versus women (Peersman et al., 2011; Mohammad and Yang, 2011; Bamman et al., 2012; Schwartz et al., 2013; Bamman et al., 2014; Prabhakaran et al., 2014).”
Adhering, Steering, and Queering: Treatment of Gender in Natural Language Generation (2020, methodology): “Using such an approach on social media corpuses, for instance, Bamman et al. [3], found that female authors use more pronouns, emotional terms (sad, love, etc.), and emoti-cons, while male authors use more proper nouns, swearing, and taboo words.”
Analyzing Biases in Human Perception of User Age and Gender from Text (2016, background): “Bamman et al. (2014) also examine individuals whose language does not match their automatically predicted gender.”
Confounds and Consequences in Geotagged Twitter Data (2015, background): “5But see Bamman et al. (2014) for a much more detailed discussion of gender and standardness.”
Bleaching Text: Abstract Features for Cross-lingual Gender Prediction (2018, background): “Author proﬁling is the task of discovering latent user attributes disclosed through text, such as gender, age, personality, income, location and occupation (Rao et al., 2010; Burger et al., 2011; Feng et al., 2012; Jurgens, 2013; Bamman et al., 2014; Plank and Hovy, 2015; Flekova et al., 2016).”