The widespread use of social media in health communication makes it important to understand how the media’s characteristics impact health communication effectiveness. This dissertation used social norms theory to explain the persuasive impact of web content feedback system—a unique feature of social media—on people’s responses to health promotion messages posted on a social media site YouTube. Three common social media content feedback cues (comments, aggregate ratings, and message view count) were examined. These cues were expected to influence people’s health attitudes and behavioral intentions via the mediation of three types of perceived social norms (descriptive norms, injunctive norms and subjective norms).

Two experiments examined the norms-mediated model in three health contexts (smoking, binge drinking, and texting while driving). Experiment 1 examined the influence of proportion of positive comments (large, medium, vs. small) and comment focus (message vs. behavior). As expected, proportion of positive comments was negatively related to people’s perceived social approval of smoking and texting while driving (injunctive norms). However, it had a concave downward relationship with perceived social approval of binge drinking (i.e., injunctive norms). The results also suggested an important impact of comment focus on people’s responses to health promotion messages. Experiment 2 examined the influence of proportion of thumbs-up (large, medium, vs. small) and message view count (high, medium vs. low). It was hypothesized that greater proportion of thumbs-up would lead to less favorable beliefs about problem behaviors. This hypothesis was supported only when the outcome variable was texting while driving intention. View count positively predicted people’s smoking and binge drinking intentions and marginally, positively predicted texting while driving subjective norms. It had a concave downward relationship with binge drinking attitudes. Experiment 2 also found complex joint effects of view count and proportion of thumbs-up.

In spite of many findings about the influence of comments, ratings, and view count on norms, attitudes, or intentions, both study 1 and study 2 provided limited support for the hypothesized norms-based mediation. Limitations, theoretical and practical implications, and directions for future research are discussed.