Diffusion of Morality and Emotion on Twitter
Many of us would like to believe that the diffusion and acceptance of an idea or moral is based solely on its strength of structure and sound logic. But alas, this is a naive notion; like it or not, it is human nature that emotions will likely be involved in almost all forms of discourse. According to William J. Brady and his fellow researchers at New York University, this is especially true in the realm of Twitter. Brady and his team used “previously validated [separate] dictionaries to count the frequency of moral and emotional words in each tweet”. They were paying specific attention to words found in both the moral and emotional dictionary, seeking to understand the role moral and emotional words played in conversations like gun control, same-sex marriage, and climate change. Brady and his team found that each moral-emotional word used increased a tweet’s retweets by around 20%. Some of the most effective words include attack, hate, blame, and peace. Perhaps even more interesting and relevant to our class, the research team found that the diffusion of tweets with moral-emotional words had a much higher transmission rate within similarly minded groups than between groups. For instance, the graph below displays the interactions between liberals and conservatives on Twitter when discussing gun control, same-sex marriage, and climate change. Each dot is a user, and each edge represents a retweet.
When observing the interaction between two groups, each interaction with someone in the same group can be seen as a strong or positive connection, and interactions across groups can be seen as a weak or negative connection. Just as discussed in class, the formation of two segregated groups is the only way for the network to be stable. While it seems like this structure is unsustainable, it’s fascinating that it is actually proven to be the only stable possibility. I find it both extremely interesting and saddening that the default network structure for political discourse between massive amounts of people is having two groups that are essentially echo chambers.
The study discussed can be downloaded from the link below:
https://www.researchgate.net/publication/317947723_Emotion_shapes_the_diffusion_of_moralized_content_in_social_networks