UA Little Rock COSMOS Study Shows How Symbols Drive Social Media Engagement
A study from the Collaboratorium or Social Media and Online Behavioral Studies (COSMOS) Research Center at the University of Arkansas at Little Rock reveals that the use of symbols in Instagram posts significantly boost engagement and help information spread more quickly across social media platforms.
The research was led by Dr. Nitin Agarwal, founding director of COSMOS Research Center and the Jerry L. Maulden-Entergy Endowed Chair and Donaghey Distinguished Professor of Information Science at UA Little Rock, and included graduate research assistants Mayor Inna Gurung and Monoarul Bhuiyan. The study was supported through research grants funded by the National Science Foundation and the U.S. Department of Defense.
Using AI-based computer vision tools, the researchers analyzed 3,097 Instagram images, quantifying symbolic content and measuring how visual and textual coherence affected engagement. Their results showed a clear pattern: posts with more symbolic elements and better alignment between captions and imagery saw significantly higher engagement rates. These posts also spread more quickly, a dynamic demonstrated through epidemiological modeling.
“This study underscores the importance of AI and multimodal analysis in understanding how online narratives evolve and spread,” Agarwal said. “Our findings show that symbolic content and narrative coherence aren’t just aesthetic choices—they’re key factors that drive audience attention and influence.”
Presented at the 58th Annual Hawaii International Conference on System Sciences (HICSS) held in Hawaii, the study titled, “How does Semiotics Influence Social Media Engagement in Information Campaigns?”, combines artificial intelligence and socio-computational modeling to examine the impact of visual symbols in multimedia-rich posts.
The team used state-of-the-art AI models, including large language models, to decode and classify visual symbols and measure how closely captions reflected image content. This approach made the research more scalable and efficient than traditional methods.
“Research studies like this pave the way for innovation by showcasing the transformative potential of AI in deepening our understanding of visual narratives,” Agarwal added. “Moreover, the socio-computational methodologies developed through this work provide powerful tools for evaluating the influence of multimedia-rich information campaigns and enhancing strategic communications.”
The findings could have broad applications in areas such as public health, strategic communication, national security, and additional areas where information campaigns seek to reach, engage, and influence audiences online.