The spread of personal computers and Internet empowers ordinary people and gives birth to social media. Social media is the collection of Internet-based applications that allow the creation and exchange of user-generated content. Unlike traditional production of information content, the social media production depends on social interactions. In my thesis, I study the effects of online social interactions between users on their content contribution behavior. There are three studies. In the first study, I focus on the social influence in online product ratings. Based on a quasi-experiment design that leverages the observation of the evolution of the friendship network, my findings strongly support the existence of social influence in users’ rating behavior. This social influence g...[ Read more ]

The spread of personal computers and Internet empowers ordinary people and gives birth to social media. Social media is the collection of Internet-based applications that allow the creation and exchange of user-generated content. Unlike traditional production of information content, the social media production depends on social interactions. In my thesis, I study the effects of online social interactions between users on their content contribution behavior. There are three studies. In the first study, I focus on the social influence in online product ratings. Based on a quasi-experiment design that leverages the observation of the evolution of the friendship network, my findings strongly support the existence of social influence in users’ rating behavior. This social influence generates biases in online review systems. My results further reveal that the significance of social influence varies across books and users. In the second study, I investigate the effect of friend role identity on user participation using a natural experiment resulting from a change in the social networking function of the site that I study. My findings suggest that making friend identity salient increases user participation both directly and indirectly through network effects. Further, friend identity shifts the focus of a user’s online activities. In the third study, I focus on the text review contribution and empirically differentiate between normative social influence and informational social influence by differentiating between groups of online social contacts in the empirical model. The results suggest that, on one hand, normative social influence is positive between online friends; and on the other hand, informational social influence introduce a crowd-out effect (negative social influence). That is, review contribution from the followees of a user reduces her review contribution. Overall, my thesis fills several gaps in the empirical research about social media and online social interactions.