The effect of college students' attachment anxiety, social media use to avoid negative emotions, and perceived stress from COVID-19 on depression

Objectives The purpose of this study is to test the mediating effect of perceived stress due to COVID-19, using social media to avoid negative emotions in the process of influencing depression of college students' attachment anxiety. Methods To this end, data from 279 male and female college students in Korea were collected through an online survey from October 15 to October 25, 2020, SPSS 20.0 was used for technical statistics and correlation analysis, and Mplus 7.0 program was used for structural model and mediating effect verification. Results The main results of this study are as follows. First, the attachment anxiety of college students showed a significant positive correlation with the use of social media to avoid negative emotions, perceived stress and de-pression caused by COVID-19. Second, the mediating effect of using social media to avoid negative emotions was confirmed in the relationship between attachment anxiety and depression. Third, the mediating effect of per-ceived stress due to COVID-19 was confirmed in the relationship between attachment anxiety and depression. Finally, it was found that the double mediating effect of social media use to avoid perceived stress and negative emotions caused by COVID-19 was significant in the effect of attachment anxiety of college students on depression. Conclusions Taken together, it was found that college students with attachment anxiety perceived stress caused by COVID-19 higher. In addition, they used social media as a way to avoid negative emotions, which affected depression. This study is meaningful in that it revealed the link between social media use and depression among young people and the psychological mechanism that affects depression among college students in the COVID-19 situation in a situation where face-to-face activities have been reduced and non-face-to-face interpersonal rela-tionships have been recommended.