Beyond self-reports of social media exposure and engagement: The feasibility of integrating social media profiles into a longitudinal vaping-related study
Dhiraj Murthy, Caroline North, Srishty Arora, Jackson Sorenson, Qinghua Yang, Bara S. Bataineh, Seth J. Schwartz, Alexandra Loukas, Anna V. WilkinsonObjective
Engagement with tobacco-related content on social media is associated with increased risk for tobacco use in young adulthood. However, knowledge about effective strategies for collecting tobacco-related social media data from samples of young adults participating in online survey research is limited. In this study, we evaluated the feasibility of linking survey responses with verified social media profiles and using AI-assisted methods to identify vaping-related content in an ongoing, web-based longitudinal tobacco/nicotine study.
Methods
We collected validated handles from 504 (out of 1,114 total) Mexican American college students. Specifically, we scraped public data from Reddit (90 users, 893 posts, and 6,057 comments), X (98 users and 28,953 posts), Instagram (253 users, 2,111 posts, 1,381 comments, and 815 reels), and YouTube (263 users, 1,000 videos, and 314 video transcripts).
Results
Using ChatGPT and manual review, we identified a limited amount of smoking-related posts/comments (
Conclusions
Nonetheless, our study confirms the feasibility of: profile linkage, social media data retrieval, and AI-assisted content identification from a longitudinal sample of college students via online survey. Findings also provide a greater sense of specificity regarding the types of social media data that can be collected and analyzed. However, due to the small number of tobacco-related posts identified among our sample, future surveillance studies should focus on oversampling tobacco users to increase the likelihood of capturing relevant social media content.