Summary: New research shows that small shifts in a person’s writing style reflect which social group is guiding their behaviour at a given moment. The study finds people adjust how they write to fit the norms of the group identity that is salient, and these shifts can be detected algorithmically.
Source: Lancaster University
Small but consistent changes in writing can reveal which social group a person is aligned with at a given moment, say psychologists.
Group memberships are central to human identity. Most people belong to multiple groups—ranging from local clubs and online communities to family roles and national identity—and these memberships often shape how we think, feel and act in different situations.
When a particular membership becomes relevant, people tend to adopt the norms associated with that group so their behaviour appears appropriate for the situation. The new study, conducted by researchers at Lancaster University, the University of Exeter, Imperial College London and University College London, demonstrates that these normative pressures are visible not only in behaviour but in linguistic style.
The research shows that subtle stylistic features of text—rather than topic words—carry signals about which group identity is active. By analysing patterns such as pronoun use, expressions of cognition and emotional language, the team developed a tool that can infer which of two group identities influenced a given piece of writing with roughly 70% accuracy.
The paper, published in Behavior Research Methods, is titled “ASIA: Automated Social Identity Assessment using linguistic style.” Professor Mark Levine of Lancaster University commented: “Our work shows that changes in how people view themselves can be detected from their writing. These identity traces remain even after accounting for the subject matter and the platform used. The ASIA toolkit advances our ability to study how psychological identity shapes behaviour.”
To validate the method, the researchers focused on two specific identities—being a parent and being a feminist—and examined anonymous online forum posts from platforms such as Reddit, Mumsnet and Netmums. The study tracked how the same authors shifted their linguistic style when moving between contexts where one identity or the other was salient.
“People are not fixed; we shift between different identities depending on the situation,” said Dr Miriam Koschate-Reis from the University of Exeter. “Right now, many individuals must move between identities—for example, between parent and employee—as they juggle childcare, homeschooling and work. These shifts affect behaviour in multiple ways. In our study, we tracked active identities by analysing language use.”
The researchers deliberately excluded obvious “content” words (such as ‘childcare’) to focus on stylistic markers: pronouns, cognitive-processing terms, markers of emotion, and other function words that reflect how people frame and present themselves. Those patterns shifted in line with the identity participants were prioritising; for instance, when asked to think of themselves as a parent in an experiment, participants’ language showed predictable changes consistent with that identity.
Beyond advancing basic science, the team highlights practical applications. “We are currently focusing on mental health,” Dr Koschate-Reis said. “ASIA is the first method that enables researchers to measure how people access different group identities at scale, using naturally occurring online text. This opens the door to studying identity acquisition—for example, how someone becomes a first-time parent—and whether difficulties in adopting a new identity contribute to problems such as postnatal depression or anxiety.”
The authors suggest ASIA could inform policies and interventions that support identity transitions, including targeted mental health support. Group identities influence thoughts, emotions and behaviours across many domains—from workplaces and educational settings to political activism—so a tool that detects identity salience in text has broad potential.
Research continues into how much voluntary control people have over switching identities and which cues in the environment trigger those switches. Dr Koschate-Reis notes that manipulating contextual cues—such as working in a library rather than a coffee shop—might make it easier to adopt an “academic” identity and its associated writing style.
Funding: The study was funded by the Engineering and Physical Sciences Research Council (EPSRC).
About this writing and psychology research news
Source: University of Lancaster
Contact: Gillian Whitworth – University of Lancaster
Image: The image is in the public domain
Original Research: Open access.
“ASIA: Automated Social Identity Assessment using linguistic style” by Mark Levine et al. Behavior Research Methods
Abstract
ASIA: Automated Social Identity Assessment using linguistic style
Our various group and category memberships are central to identity and influence cognition, emotion, behaviour and social relationships across many contexts. They are also linked to mental and physical well-being. Yet questions remain about how different memberships interact and how people come to acquire them cognitively and emotionally, especially outside laboratory settings.
Existing assessment methods have been limited in their ability to analyse naturally occurring data—such as online interactions—to study the dynamics of group membership in everyday life. To address this gap, we developed ASIA (Automated Social Identity Assessment), an analytical protocol that uses linguistic style features to infer which group membership is salient at a given moment. An open-source Jupyter Notebook tutorial accompanies the method to support reproducibility and wider use.
In this paper we outline the challenges in studying salient group memberships and demonstrate how ASIA can address some of these challenges. We apply the protocol to distinguish which of two specific memberships—parents versus feminists—is salient in online forum posts, and we test measurement validity across additional corpora and an experimental study. We conclude by discussing future directions for research and methodological development.