Positive and Negative Sentiment in Social Media Direct Messages Predicts Negative Emotion Differentiation Among Adolescents.

Abstract

Negative emotion differentiation characterizes the ability to draw distinctions between discrete negative emotional experiences with high specificity. Negative emotion differentiation is linked to improved emotion regulation and may be a key marker of adaptive emotional functioning. The present study explores how emotional language used by adolescents in daily life relates to their ability to distinguish among discrete emotions using two linguistic measures of emotion. Adolescents (N = 53; 28 girls, 23 boys, 2 outside gender binary; 73.7% non-White) rated their current negative emotions (e.g., anxious, fearful, lonely) via ecological momentary assessments (EMA) three times a day for 2 weeks. They also shared 94,497 of their direct messages sent on Instagram, one of the most popular social media platforms among adolescents. From these measures, we respectively computed participants' degree of negative emotion differentiation across the 2 weeks and the positive, negative, and neutral sentiment of direct messages using a dictionary-based sentiment analysis (VADER). Results reveal that adolescents with higher negative emotion differentiation also had a greater percentage of positive and negative valenced direct messages. These findings are consistent with the notion that individuals greater in emotion differentiation experience and express a broader range of emotions in daily life.