In many online social systems, social ties between users play an important role in dictating users' behavior. One of the ways this can happen is through social influence, the phenomenon that the actions of a user can induce his/her friends to behave in a similar way. In systems where social influence exists, ideas, modes of behavior, or new technologies can diffuse through the network like an epidemic. Therefore, identifying and understanding social influence is of tremendous interest from both an analysis (e.g., predicting the future of the system) and a design (e.g., designing viral marketing strategies) point of view.

In this talk, I will give a general overview of models for diffusion in social network, and then discuss the problem of identifying social influence in the data. This is a difficult task in general, since there are many other factors such as homophily or unobserved confounding variables that can induce statistical correlation between the actions of friends in a social network. Thus, distinguishing influence from those other factors is essentially the problem of distinguishing correlation from causality, a notoriously hard problem. Despite this, I will show how in an environment where the time stamp of the actions are observable, we can design simple statistical tests that distinguish between models of social influence and those that replicate the aforementioned sources of social correlation. I will sketch the proof of a theoretical justification of one of the tests, and present simulation results on randomly generated data and real tagging data from Flickr. The results exhibit that while there is significant social correlation in tagging behavior on this system, this correlation cannot be attributed to social influence.