We would like to invite you to participate in Management Engineering(ME) Seminar.

1. When: April 3rd (Tuesday), 16:00 ~ 17:20
2. Where: Room 9708
3. Speaker: Prof. Tianshu Sun (University of Southern California)
4. Topic: Displaying Things in Common to Encourage Friendship Formation:  A Large Randomized Field Experiment
5. Research field: IT Management

Abstract
Friendship formation among people with different background is of central importance to online social networks sites and the society in general. In this study, we investigate and empirically test whether and how displaying things in common may encourage friendship formation. In collaboration with a large online social network, we carefully design and implement a large-scale in vivo randomized field experiment involving over 50 million viewer-profile pairs, in which the prominence of ‘things in common’ information is randomly varied when an user is browsing another’s profile. We find that displaying things in common may significantly increase the probability of friendship formation. More interestingly and surprisingly, displaying things in common is only effective for people who have little in common: the causal effect is large and significant for those viewer-profile pairs with no mutual friends and only share few things in common. Such findings suggest that our network intervention may effectively connect people with different background, increase diversity of users' friendship ties and reduce homophily in the social network. We further explore the causal effect of different types of things in common (mutual city, work, hometown, education and page likes) for a wide range of users with different demographics and social status. The discovered heterogeneity further sheds light on the underlying mechanism. In summary, our experiment is among the first to demonstrate the effectiveness of network intervention in encouraging friending behavior, especially for people who share little in common. Our study also contributes to the network theory by provide a clear empirical evidence and nuanced insights on the role of preference (over similar attributes) in driving network formation and homophily.