Related Experiment Video
Updated: Jan 4, 2026

Integrating Computerized Linguistic and Social Network Analyses to Capture Addiction Recovery Capital in an Online Community
Published on: May 31, 2019
Exploring the association between problem drinking and language use on Facebook in young adults
Davide Marengo1, Danny Azucar2, Fabrizia Giannotta3
1Department of Psychology, University of Turin, Turin, Italy.
Social media language use, including words about family and positive emotions, is linked to lower alcohol misuse. Conversely, terms related to nightlife and coarse language indicate higher problematic drinking, with open-vocabulary analysis offering better prediction.
Area of Science:
- Computational linguistics
- Psychology
- Public health
Background:
- Social media text variations correlate with user demographics, psychosocial factors, and behaviors.
- Problematic alcohol use is a significant public health concern with potential links to online communication patterns.
Purpose of the Study:
- To investigate the association between Facebook language use and problematic alcohol consumption.
- To compare the predictive power of closed-vocabulary (LIWC) and open-vocabulary (LDA) text analysis methods for alcohol use.
- To explore the utility of machine learning in identifying linguistic markers of problematic drinking.
Main Methods:
- Collected Facebook texts from 296 adult social media users.
- Analyzed text using Linguistic Inquiry Word Count (LIWC) and Latent Dirichlet Allocation (LDA).
- Assessed alcohol use via the AUDIT-C questionnaire and employed Random Forest for predictive modeling.
Main Results:
- Negative associations found between words related to family, school, and positive emotions and problematic alcohol use.
- Increased frequency of words concerning sports, politics, nightlife, and coarse language among problematic drinkers.
- Open-vocabulary features (LDA) demonstrated superior predictive accuracy (r = .46) compared to closed-vocabulary features (LIWC, r = .28) for AUDIT-C scores.
Conclusions:
- Specific language patterns on social media are associated with problematic alcohol use.
- Open-vocabulary text analysis provides richer insights than closed-vocabulary methods for predicting alcohol consumption behaviors.
- Online text analysis holds potential for developing novel alcohol screening tools.
Related Concept Videos
Self-Presentation: Self-Monitoring and Self-Handicapping
Substance Use Disorders Affecting Sleep
Understanding the concepts of physical dependence,...
Language and Cognition
Stress Prevention and Stress Management Techniques IV
CNS Depressants: Alcohol and Nicotine
Social Anxiety Disorder

