Spread of misinformation on Facebook
This article discusses a study that analyzed over 1300 Facebook pro or anti-vaccine pages with nearly 100 million followers to produce a network map of how information spread. The study found that though anti-vaccine pages have fewer followers than pro-vaccine ones, they grow faster and connected to more undecided pages.
\
This study is relevant to our class, as it discusses the spread of misinformation on Facebook using a network graph. The graph leads to several interesting observations. First of all, the network comprises a single giant component. There doesn’t seem to be a node that isn’t connected to at least one other node in the graph. However, there are two clusters of connected components within the network, one in the upper half mostly dominated by pro-vaccine pages and undecided pages, and the other in the bottom half mostly dominated by anti-vaccine and undecided pages. Each cluster has highly embedded nodes with an extremely short distance between nodes. Moreover, the distance between nodes seems to increase towards the middle of the graph. Nodes in the upper half connected to a node in the lower half are most likely to be local bridges connecting the two cluster groups.
Based on the study, experts also suggested that pro-vaccine pages tend to be in an echo-chamber. It’s interesting to think about how an ‘echo chamber’ can be described using the concepts learned in this class. Such a network will probably have deeply embedded nodes, with fewer nodes functioning as local bridges. Due to the high levels of embeddedness, there will be greater trust between the nodes. All relationships within the echo chamber network will be positive while all relations to nodes outside the network will be negative. Therefore, echo chambers reflect a balanced network.
It would be interesting to see what this network would look like if it had taken into account the strength of ties and whether connections are positive or negative. There is a limit to how much can be inferred from a network graph, only depicting connections and the pathways between nodes. It should be noted that social networks are continually changing and that this graph should only be taken as a snapshot of the network at a given time.