Objective: This study investigated the SNS activities of Japanese university students to clarify the features of depressive symptoms from patterns of SNS postings.
Method: College students utilizing Twitter (n=158, 94 males, 63 females, 1 unknown; average age=18.89, SD=0.90, valid response rate=73.8%) were grouped into depressive categories: normal (n=57), mild (n=75), and moderate or more (n=26), based on Self-Rating Depression Scale (SDS) scores. Twitter activity data was collected for one month from the participants' Twitter accounts to explore difference between groups.
Result and Discussion: Group differences in Twitter activity data revealed higher rates of original tweets (monologues) in the morning in the mild and moderate or more groups. Text mining was applied to 1,919 original tweets posted in the morning, and the relationship between depressive category and extracted words was examined by correspondence analysis. Two components— “busy in real life” and “escape from real life”—were obtained. Further, drawing from the diagram of correspondence analysis, the mild group tended to express realities of their daily situation such as busyness with academic activities, while the moderate or more group tended towards expression of escapist attitudes from present reality and features that could be regarded as manic defense. In future, expanding the range of targeted posts should be helpful to determine if the findings obtained in this study hold true for all segments of the day.
View full abstract