Emotion shapes the diffusion of moralized content in social networks

Abstract

Political debate concerning moralized issues is increasingly common in online social networks. However, moral psychology has yet to incorporate the study of social networks to investigate processes by which some moral ideas spread more rapidly or broadly than others. Here, we show that the expression of moral emotion is key for the spread of moral and political ideas in online social networks, a process we call "moral contagion." Using a large sample of social media communications about three polarizing moral/political issues (n = 563,312), we observed that the presence of moral-emotional words in messages increased their diffusion by a factor of 20% for each additional word. Furthermore, we found that moral contagion was bounded by group membership; moral-emotional language increased diffusion more strongly within liberal and conservative networks, and less between them. Our results highlight the importance of emotion in the social transmission of moral ideas and also demonstrate the utility of social network methods for studying morality. These findings offer insights into how people are exposed to moral and political ideas through social networks, thus expanding models of social influence and group polarization as people become increasingly immersed in social media networks.

Keywords: emotion; morality; politics; social media; social networks.

Fig. 1.

Moral-emotional language predicts the greatest number of retweets. The graph depicts the number of retweets, at the mean level of continuous and effects-coded covariates, predicted for a given tweet as a function of moral and moral-emotional language present in the tweet. Bands reflect 95% CIs. An increase in moral-emotional language predicted large increases in retweet counts in the domain of (A) gun control, (B) same-sex marriage, and (C) climate change after adjusting for the effects of distinctly moral and distinctly emotional language and covariates.

Fig. 2.

The effect of moral contagion is greatest within political in-groups compared with political out-groups. The graph depicts expected retweet count as a function in-group/out-group network and moral-emotional content. Bands reflect 95% CIs. Moral-emotional language was associated with a significantly larger retweet rate for the political in-group for the topics of (A) gun control and (C) climate change. For the topic of (B) same-sex marriage, the result was not significant, although in a consistent direction.

Fig. 3.

Network graph of moral contagion shaded by political ideology. The graph represents a depiction of messages containing moral and emotional language, and their retweet activity, across all political topics (gun control, same-sex marriage, climate change). Nodes represent a user who sent a message, and edges (lines) represent a user retweeting another user. The two large communities were shaded based on the mean ideology of each respective community (blue represents a liberal mean, red represents a conservative mean).