California Accepted Papers Paper:
Does Social Media Use Predict Moderate and Severe Anxiety? Evidence from High School Students
*Names in bold indicate Presenter
Evidence from High School Students
John Keefer, Cal Poly
Jacqueline Doremus, Cal Poly
Over 14% of adolescents in the United States experience depressive symptoms, a number up over 37% since 2005. Additionally, over 32% of high school students experience anxiety symptoms, with the number seeking professional mental health services doubling since 1980. The increased rates may be due to the larger number of students willing to report symptoms or seek mental health care, with a shift in society’s view towards those who struggle with mental illness. However, there are growing concerns that the rise of social media use may be contributing to the increases in anxiety and depression rates observed in high school teens.
We were able to hand collect data on students from four high schools in a San Francisco Bay Area school district to analyze whether social media use is correlated with anxiety and depression in teens. Our survey collected a rich set of demographic, economic, family, and lifestyle variables correlated with anxiety that we were able to control for in our models. We used linear probability, probit, and random forest models to estimate the statistical relationship while controlling for our other observed characteristics.
We find that students with the highest social media use, over two hours daily, experienced a significantly higher rate of occurrence of moderate to severe anxiety and depression. It was estimated, while controlling for a number of important variables, that those with high social media use had a 10 percentage point increase in the probability of having moderate to severe anxiety and depression. However, the relationship we found was non-linear, with lower levels of social media use not being correlated with higher instances of anxiety symptoms. The increased rate of anxiety and depression was still found when controlling for important variables such as sleep or bullying suggesting they were not the driving mechanisms. Additionally, the increased rate was still observed when controlling for demographic, family, and economic variables, providing evidence that the higher levels of anxiety and depression were not due to differences in the types of students who used higher levels of social media.
Our paper contributes to the emerging literature assessing the consequences in the dramatic change in the way teens socialize. Our study does not find a causal effect of social media use, with a key limitation being the risk of reverse-causality. It is possible that teens who experienced higher levels of anxiety or depression used more social media as a result, in which case reducing social media use would not reverse the increases in anxiety and depression. Further research is greatly needed to determine if a causal relationship exists.
