Information Cascades on Twitter
Information cascades occur in situations where individuals observe the actions of those around them and follow the behavior of preceding individuals. The actions of individuals within social media outlets, such as Twitter, make for a perfect example of this phenomenon. An information cascade in these areas can be defined as a situation where one individual observes the actions of other individuals (tweets, posts, etc) and as a result, starts sharing and reposting it.
The 140 character limit on tweets allows for concise yet informative information to be posted, and the retweet feature serves as a form of propagation amongst people within the same network. During times of emergency, typically public officials, emergency agents, and news outlets post real-time updates on Twitter. These individuals can be thought of as external sources that introduce messages into the network that reach a subset of nodes. These messages spread as nodes interact with one another. When a user is influenced by a tweet, they may retweet the message, thereby allowing their followers to see the tweet. Their followers will then evaluate the information, and can choose to either perform the action of retweeting or not. This is the general basis of how information diffuses through Twitter.
In April of 2010, an armed robbery occurred nearly one mile away from the RPI college campus at 8:15am. Hours later, the robber was spotted on campus. The RPI emergency alert system was used to notify the campus community regarding the threat. The first alert was sent at 9:30am, and was followed by two additional alerts at 10:48am and 11:48am. Data from Twitter was collected regarding these series of events, and a total of 641 tweets were found spanning over the course of the day. This dataset was analyzed and the researchers found that this activity represented an information cascade. 318 unique users were identified in this dataset. Of the 318 users, 136 users retweeted at least 1 earlier tweet and 113 users sent at least one tweet over direct message or publicly replied to a preceding user. The degree of a given user also demonstrates insights about the user themself. A high degree would infer that the user may be a reliable source that provides valuable information to other users. This can be inferred from the fact that their messages are frequently re-shared with others via retweets, direct messages/shares, and public replies. All in all, the actions of users on Twitter provide valuable insight into the manner in which individuals behave in situations where information diffuses to users in a social network.
Sources: https://www.researchgate.net/publication/239761833_Information_cascades_in_social_media_in_response_to_a_crisis_A_preliminary_model_and_a_case_study
https://reader.elsevier.com/reader/sd/pii/S1877050920315283?token=461ECFC1BB0B64F22970882C7F9326460CCED0DCA4965963676C57F9BF65A041865D9AB9E00B3E2FB597B9CB2A2F9965&originRegion=us-east-1&originCreation=20211031203852