Connecting the Dots: When Personal Information Becomes Personally Identifying on the Internet

Abstract

With online social media such as weblogs (blogs), authors seemingly control how much self-identifying information they disclose. However we find that that even authors who wish to remain anonymous will share expressive and access-enabling information which, when combined, can be used to positively identify the person. In a case study of three anonymous blogs we demonstrate how to combine investigative analysis with statistical techniques to identify anonymous authors with a high degree of accuracy. Paradoxically, anonymous authors feel as if they can be honest and open with their thoughts and opinions, and thus may be more likely to share more information than they might if their identities were known.