Research Showcase Gallery (Poster 2102)

Using Linguistic Inquiry and Word Count (LIWC) to Predict Future Alcohol-Related Risk From Free-Response Language During Simulated Alcohol Offers

Abstract

Language reveals subtle information about cognition and behavior. Computational language analysis can predict this information more simply and reliably than other methods. Language analysis has not been applied to alcohol-drinking situations, in which the language used may predict outcome risk. The aim of this study was to use language analysis software to predict risky drinking from language used during an alcohol-offer simulation. We hypothesized that individual words used in response to the simulation would correlate with risky drinking, as measured using Alcohol Use Disorder Identification Test (AUDIT) scores at baseline and 8-months following the simulation. We used Linguistic Inquiry and Word Count (LIWC) to analyze the transcripts of 60 first-year college students (59.8% female, 81.5% white) and correlated the LIWC data with the outcome data. We found significant positive correlations between AUDIT scores and language regarding social affiliation (r = .333, p < 0.05), work life (r = .316, p < 0.05), and home life (r = .458, p < 0.01), and significant negative correlations with the use of quantifiers (e.g., few, many; r = -.296, p < 0.05) and prepositions (e.g., to, with; r = -.347, p < 0.05). There were also significant positive correlations between willingness to consume alcohol and social affiliation language (r = .292, p < 0.05), and reward language (r = .307, p < 0.05). These results show individual words used in drinking situations can predict risky drinking behavior and related consequences, which supports the utility of this software in research and clinical practice.


About the Presenter

default user image

Sean Noudali

Sean Noudali is a first-year graduate student in the Experimental Psychology PhD program at Washington State University, Vancouver. He is a member of the Promoting and Treating Health lab working with Dr. Benjamin Ladd. Sean received his Bachelor of Science in biological sciences from Arizona State University in 2017. He is currently studying the relationship between cannabis use and emotion processing.