The Relationships Between Users' Negative Tweets, Topic Choices, and Subjective Well-Being in Japan

Abstract

This study examined the relationships between expressions in Tweets, topic choices, and subjective well-being among undergraduates in Japan. The authors conducted a survey with 304 college students and analyzed their Twitter posts using natural language processing (NLP). Based on those who posted over 50 tweets, the authors found that (1) users with higher levels of social skills had fewer negative tweets and higher levels of subjective well-being; (2) frequent users posted both positive and negative tweets but posted more negative than positive tweets; (3) users with fewer negative tweets or with more positive tweets had higher levels of subjective well-being; and (4) “safe” topics such as social events and personal interests had a positive correlation with the users' subjective well-being, while debatable topics such as politics and social issues had a negative correlation with the users' subjective well-being. The findings of this study provide the foundation for applying NLP to analyze the social media posts for businesses and services to understand their consumers' sentiments.