Socially-Framed Text Appeals Predict Drinking Behavior in SNS Alcohol Advertising: An Eye Tracking Study (Preprint)

<sec> <title>BACKGROUND</title> Alcohol-related marketing on social networking sites (SNS) often combines emotionally framed textual appeals with alcohol imagery, mimicking peer-to-peer communication. While regulatory efforts primarily target visual cues, less is known about the influence of emotionally charged textual content on drinking behavior. </sec> <sec> <title>OBJECTIVE</title> This study investigated how visual attention to textual versus image-based content in SNS alcohol posts relates to drinking behavior and whether effects differ by emotional message type (social, ego, or sensory appeals). </sec> <sec> <title>METHODS</title> A laboratory-based eye-tracking experiment was conducted with 157 young adults (aged 18–30). Participants viewed Instagram-style alcohol advertisements containing one of three textual appeal types paired with alcohol-related imagery. Total fixation duration (TFD) on text and image areas was recorded. Drinking frequency over the past 30 days was assessed via self-report. Multiple linear regression analysis assessed main and interaction effects of attention and message type. </sec> <sec> <title>RESULTS</title> Attention to alcohol imagery was not significantly associated with drinking frequency (β = 0.06, P = .41). However, attention to textual content interacted significantly with message type (F(2,120) = 4.35, P = .015). Greater fixation on social appeals predicted higher drinking frequency (β = 0.29, P = .01), while fixation on sensory appeals predicted lower drinking frequency (β = –0.25, P = .04). </sec> <sec> <title>CONCLUSIONS</title> Emotionally framed textual messages particularly social appeals play a critical role in shaping alcohol use behavior on SNS. Because these peer-like messages resemble authentic communication, they may evade visual-focused regulatory scrutiny. Findings highlight regulatory efforts to address emotionally engaging textual content alongside imagery. </sec>