Using Twitter Post Data to Ascertain the Sentiment of Alcohol-related Blackouts in the United States
Location
Online
Event Website
https://hicss.hawaii.edu/
Start Date
3-1-2023 12:00 AM
End Date
7-1-2023 12:00 AM
Description
Research shows variability in how alcohol-related blackouts (periods of memory loss during/after drinking) are subjectively evaluated. We accessed 3.5 million original Tweets written in the U.S. between July 2009 and February 2020 that referenced blackouts, and coded the sentiment (positive or negative) of those Tweets, using the machine learning function of a Twitter-sponsored commercial platform. The sentiment of Tweets was examined by day of week and compared to the sentiment of blackout Tweets on certain holidays to non-celebration matched days. Tweets were more likely to have a positive (73%) than negative sentiment, and positive Tweets were more common during weekends. Relative to typical non-celebratory weekends, a greater proportion of blackout Tweets were positive around Thanksgiving and New Year’s Eve, though differences were not observed relative to several other celebratory periods (e.g., Superbowl). Results have implications for online interventions, which can use social networking sites to target alcohol during high-risk periods.
Recommended Citation
Merrill, Jennifer; Riordan, Benjamin; Ward, Rose Marie; and Raubenheimer, Jacques, "Using Twitter Post Data to Ascertain the Sentiment of Alcohol-related Blackouts in the United States" (2023). Hawaii International Conference on System Sciences 2023 (HICSS-56). 5.
https://aisel.aisnet.org/hicss-56/hc/social_media/5
Using Twitter Post Data to Ascertain the Sentiment of Alcohol-related Blackouts in the United States
Online
Research shows variability in how alcohol-related blackouts (periods of memory loss during/after drinking) are subjectively evaluated. We accessed 3.5 million original Tweets written in the U.S. between July 2009 and February 2020 that referenced blackouts, and coded the sentiment (positive or negative) of those Tweets, using the machine learning function of a Twitter-sponsored commercial platform. The sentiment of Tweets was examined by day of week and compared to the sentiment of blackout Tweets on certain holidays to non-celebration matched days. Tweets were more likely to have a positive (73%) than negative sentiment, and positive Tweets were more common during weekends. Relative to typical non-celebratory weekends, a greater proportion of blackout Tweets were positive around Thanksgiving and New Year’s Eve, though differences were not observed relative to several other celebratory periods (e.g., Superbowl). Results have implications for online interventions, which can use social networking sites to target alcohol during high-risk periods.
https://aisel.aisnet.org/hicss-56/hc/social_media/5