The WeChat Supernetwork
The US government’s recent attempt to ban WeChat a few weeks ago (around September 20, 2020) stimulated my interest in researching more about WeChat. The most widely used social media platform in China, WeChat can message, video chat, audio call, and allow you to post and see your friends and family’s “Moments” (similar to Stories and Posts on Instagram). In China it can even serve extra functions such as acting as a virtual wallet, hailing taxis, etc. But as a social media platform first and foremost, it is a main source of information dissemination for the millions of its Chinese and Chinese-American WeChat users. But how exactly does it spread information so quickly and powerfully, or from a network standpoint, how does information spread from one node to the next and to other clusters of nodes in a way that makes it so powerful?
“Research on the Identification of Key Nodes in the Process of WeChat Epidemic Information Dissemination: A Supernetwork Perspective” by Peng Wu and Di Zhao, published this August 2020, analyzes the network topology and identifies key nodes at work with spreading COVID-19-related rumors and false information about the epidemic in China. Wu built a model of the WeChat supernetwork (Figures 1 and 3, which is based on the adjacency matrix from Figure 2), which takes into account 3 major different types of nodes and layers 3 different networks all interconnected with edges, thereby taking concepts from class about relatively simple network graphs and nodes to a whole new level that is more realistic given the complexity of the real world and social media. The three different types of nodes include ordinary private WeChat accounts or users (p), WeChat groups (m), and public WeChat accounts (i), all of which are major outlets of spreading information. Public accounts spread information through weak ties to their fans, and they often form a star network with their fans (Figure 1). Star networks allow the public accounts to spread information quickly, as well as far and wide (extensive reach) (Figure 3). Private accounts spread information mainly through strong ties with their friends and family, forming a star network, and relatively weaker ties with acquaintances. Due to the strong ties, users tend to believe information shared in this mode more; in other words, they found the information more credible. However, the speed of information spread is slower and its reach is severely more limited to only a circle of friends (as seen in the turquoise network in Figure 3). The WeChat group, on the other hand, combine both characteristics of how public vs private accounts spread information based on weak and strong ties. Consisting of nodes with a mixture of strong and weak ties, the WeChat group has the potential to break the tradeoff between reliability, and speed and spread. In other words, WeChat group chats allow information to spread faster and wider than private accounts, but to a lesser extent than public accounts. From
Figure 1
Figures 2 and 3
here, I can see it confirmed that weak ties tend to be where individuals receive more novel information more quickly, as we learned in class. For example, similar to why most people obtain jobs through novel information from acquaintances (weak ties) rather than friends (strong ties), new information about the epidemic spread faster and wider through public accounts and WeChat groups with their weak ties as compared to private accounts with their strong ties. It was also definitely interesting to see the interaction of different types of nodes in different layers of networks.
Another component to this study identifies the key nodes with high effectiveness in spreading epidemic-related information. Using mathematical calculations and modeling, Wu found that nodes i3, i4, and i5 have the highest degree of centrality and propagation ability (both of which are effective measures of identifying key nodes). To me this is not surprising, as public accounts with weak ties have the widest reach and can spread information quickest out of the three types of nodes. Most influential among WeChat groups comes m4, with the highest degree of centrality and thus the strongest spreading capability. Lastly, p3 and p4 have the highest degree of centrality and thus the strongest spreading capability among their circle of friends. I can see that the implications for this are large. To prevent or control the spread of rumors and false information such as those related to COVID-19, it is best to target or influence these key nodes that are the most effective at spreading information quickly to the most people. But to do that, one must be able to identify them first. Additionally, given the presence of public accounts and WeChat groups, ordinary WeChat users also cannot be underestimated.
As a Chinese-American myself, I can see firsthand the network implications of a ban on WeChat. My immigrant parents and many others in the US primarily use WeChat as their only source of contact with family and friends in China. After all, they can’t use American platforms such as Facebook or Google as they are banned in China, so WeChat provides the perfect platform since it is already so widely used in China. Additionally, given the recent pandemic and resulting xenophobia, WeChat’s ability to quickly spread information far and wide across many different types of nodes and networks is a saving grace to many struggling Chinese businesses, who depend even more on fellow Chinese-Americans to help them survive. Since WeChat keeps the Chinese and Chinese-American community connected, it provides small Chinese businesses and restaurants with the most effective way to advertise and disseminate information about themselves to as many people in their given networks and beyond.
Source: Wu, Peng and Zhao, Di, “Research on the Identification of Key Nodes in the Process of WeChat Epidemic Information Dissemination: A Supernetwork Perspective”, Mathematical Problems in Engineering, vol. 2020, Article ID 6751686, 10 pages, 2020. https://doi.org/10.1155/2020/6751686