Do happy people get more from their friends?: What we can learn about social networks from text analysis

Kristina Lerman
University of Southern California (USC)

Social relationships confer numerous advantages, including delivering novel information. Although a host of socioeconomic and cognitive factors are thought to affect social relationships, few of them were empirically validated. To address this gap, we analyzed social media texts posted from within a major US metropolitan area. We measured the structure of social interactions, emotions people expressed in their messages, and linked these to the socioeconomic characteristics of places from which messages originated. Places where people engage less frequently but with diverse social contacts have happier, more positive messages posted from them, and also have better educated, more affluent residents. In contrast, sadder messages were posted from places with stronger, more frequent social interactions, but were associated with less educated and larger households. Additionally, we investigated the interplay between people’s position within their social media network and the novelty of information they received from their social contacts. Network positions linking people to a diverse set of otherwise unconnected contacts exposed them to more novel and diverse information. This suggests that networks may promote inequities, wherein the happier, more affluent, and better educated people are in network positions that potentially allow them greater access to novel information, which can be strategically leveraged.


Back to Workshop IV: Mathematical Analysis of Cultural Expressive Forms: Text Data