Twitter and Strong Triadic Closure
http://research.microsoft.com/pubs/133158/GolderYardiSocComp.pdf
In the above article, the researchers conducted an experiment where the testers were asked to rate their interest in forming to ties to other Twitter users, independently of existing connections between said users. Instead of using a conventional network where two-way edges exist between users, the article covered directed triadic closure; a network where only one-way edges exist between users. The nature of Twitter is such that any user A can follow a user B, but that does not guarantee nor imply that user B will follow user A. With a network like Facebook, two-way edges exist between “friends”; two friends C and D see each other’s posts and activities on their respective news feeds. With Twitter, however, user A will see user B’s activities, but B will not see A’s activities unless he or she chooses to follow A. The article also details the differences between symmetric and asymmetric ties: when a tie is symmetric, the tie between two users X and Y either exists or does not exist. If it is asymmetric, it is the result of directed one-step and two-step paths, that is, networks that involve two or three people. Triadic closure is absent on Twitter when a user G has followers with different opinions or thoughts on the same topic (e.g. politics, a sports team, a video game.) If G posts a favorable review of a video game, half of his followers (H) may agree with the review while the other half (I) may vehemently disagree with it. A ++- relationship exists here; two mutual ties exist between G and H, and G and I, but a negative relationship exists between H and I, violating triadic closure. Conversely, a user who posts a bad review of a book that two users thought was good would create a +– relationship: the two users have a strong tie as a result of their both thinking the book was good, and they both have weak bonds with the reviewer.
This is relevant to our previous discussions of Strong Triadic Closure in that it involves both two-way and directed networks and invokes our discussions of triadic closure and “strong” and “weak” users. Strong Triadic Closure is present in relationships that are both one-way and two-way between several users, and can often benefit everyone involved as we discussed. A Twitter user who has more followers will dominate a given network in that he will have control over what content gets posted and what information is delivered. These power users will often influence the decisions and options of users involved in the network, and make it difficult for them to go elsewhere in the network for information.