I feel you: Mixed-methods study of social support of loneliness on twitter

Abstract

Loneliness is considered an epidemic in the United States due to its widespread and harmful effects to psychological and physiological well-being. Twitter provides the option of anonymity, a large audience and a space where feelings of loneliness can be expressed, and feedback received. In this mixed-methods study, based on a sample of 4 million tweets containing expressions of loneliness, we examine factors associated with eliciting feedback and types of possible social support therein. We examine feedback both quantitatively in terms of number of likes, retweets, and replies, and qualitatively by annotating its content. We apply the categorization of social support and test the applicability of concepts of directedness, person-centeredness and invisible support to a sample of replies. Supporting previous literature, we show that Twitter users with larger social networks and those who use a more positive language are more likely to receive feedback, conversely swearing is associated with fewer responses. Most common social support provided is emotional, followed by esteem and information support, all of which often include the elements of invisible support including smileys, images, and text formatting. However, there is a fraction of replies which may be considered online bullying, pointing to avenues of possible needs for intervention.