Insights into adolescent well-being from computerised analysis of written language

Abstract

Aim: To examine associations between patterns of language use and early adolescent well-being.

Methods: Participants were 1763 Australian 11- to 12-year-olds in the Child Health CheckPoint. Six patterns of language use were identified from a writing activity using Linguistic Inquiry and Word Count and factor analysis: Acting in the present and future, Positive emotion, Gender and relationships, Self-aware, Inquisitive and time focused, and Confident. Well-being measures represented a spectrum from negatively to positively framed psychosocial health. Associations between language use and well-being were estimated using linear regression adjusted for age, sex and social disadvantage.

Results: Positive emotion (high emotional tone, positive emotion) was associated with better general well-being (standardised regression coefficient (SRC) 0.05; 95% confidence interval 0.00 to 0.11; p = 0.04), life satisfaction (0.06; 0.01 to 0.11; p = 0.03), psychosocial health (0.07; 0.02 to 0.12; p = 0.01) and quality of life (QoL) (0.06; 0.01 to 0.11; p = 0.02). Similarly, Self-aware (high first person singular pronouns, authentic, low clout) was associated with better general well-being, life satisfaction and psychosocial health (SRC 0.05, 0.09, 0.08), but Confident (high clout, first person plural pronouns, affiliation) was associated with worse life satisfaction, psychosocial health and QoL (SRC -0.06, -0.09, -0.06).

Conclusion: If replicated in 'real-world' settings (e.g., social media), language patterns could provide naturalistic insights into early adolescents' well-being.

Keywords: automated computerised analysis; language patterns; linguistic analysis; mental health; well-being.