As investigators continue efforts to understand the motivations and predictive factors associated with college student drinking behaviors, they are faced with the challenge of staying up to date concerning the technology students use and how it affects their likelihood to consume alcohol.  Previous studies have found that encountering alcohol-related content on various social media platforms leads to greater alcohol consumption among college students, but little research has been performed regarding their interpersonal communications as it relates to the subject matter.  Using a small-scale data mining approach, this study sought to determine if alcohol related content in text messaging (a construct they have named “alcohol talk”) can be associated with different drinking patterns. 

The first part of the study involved researchers compiling and verifying an extensive dictionary of alcohol-related words to use for their data analyses, with 524 terms being finalized into a list called “alcohol talk”.  Next students from a large southeastern university were recruited to volunteer their text messaging data over a two-week timespan, with 267 being included in the final sample after meeting these primary eligibility criteria: current enrollment, smart phone usage, and alcohol consumption in the past year (N=267, 59.2% female).  After being entered into the study for two weeks, the participants consented to provide their text message information to the research team which was subsequently downloaded and analyzed.  Specifically, the team analyzed the frequency of alcohol talk terms used in the two-week volume of information, as well as various timing measures for when term occurred.  Lastly, the students provided their individual drinking history over the respective two-weeks, as well as their perceptions of their peers regarding drinking and substance abuse participation.  Measures of the text messaging data were computed using the Linguistics Inquiry and Word Count (LIWC) program.   

Results of the analysis found that only about 0.3% of all words in the participants’ messages were considered alcohol-talk words, with the average student using roughly 25 per week.   Despite these ostensibly small values, all but six students in the sample (97.8%) engaged in alcohol talk at some point in the two-week timespan.  Alcohol talk was found to occur more frequently in the evening, and an increase in alcohol talk word count was associated with an increase in reported drinking during that respective day (22% increase per word).  Also, the reception of alcohol talk words was associated with greater reported substance use norms among the individual’s peers. 

Takeaway: modern mobile communications between college students can be a medium through which alcohol consumption behaviors may be reliably analyzed.  Interventions exploring this avenue may be able to identify individuals in need of alcohol-related treatment but must be rigorous in maintaining the desired privacy of students. 

Jensen M, Hussong AM. Text message content as a window into college student drinking: Development and initial validation of a dictionary of “alcohol-talk.” International Journal of Behavioral Development. 2021;45(1):3-10. doi:10.1177/0165025419889175