Journal: Nature

Publishing date: January 7, 2022

Authors: Koustuv Saha et.al

Social media data and discussions can be analysed using natural language processing and artificial intelligence to provide insight into the mental health status and treatment needs of college students. 

Why does this study matter?

Social media plays a very significant role in our daily lives. Individuals express their personal life and emotions on various social media platforms. This study revealed that social media data can improve our understanding of college students' mental health, particularly their mental health treatment needs. 

Study Design

Kostov Saha et al obtained ground-truth data from on-campus mental health consultations of a large U.S. public university between 2011–2016 and collected 66,000 posts from the university’s Reddit community. Using adopted machine learning and natural language methodologies they measured symptomatic mental health expressions of depression, anxiety, stress, suicidal ideation, and psychosis on the social media data. Seasonal autoregressive integrated moving average (SARIMA) models of forecasting on-campus mental health consultations showed that incorporating social media data led to predictions with r = 0.86 and SMAPE = 13.30 (symmetric mean absolute per cent error), outperforming models without social media data by 41%.

Results and Conclusion

Language analyses revealed that social media discussions during months of high mental health visits tend to show a greater prevalence of words related to academics, academic examinations, career, and psycholinguistic attributes indicative of worse mental wellbeing, whereas the months of lower mental health visits show a greater prevalence of words related to social, partying, leisure, and psycholinguistic attributes indicative of better mental health.

Social media data can function as a “verbal sensor” to assess the mental health needs and demands of college students. It can be used to assess campus morale, regarding the mental well-being of students which can help college stakeholders, including administrators, policymakers, and wellbeing councils, to gauge the needs of the students and accordingly ensure that adequate resources are available and measures are taken to meet the demands of mental health-related services.

Another application of this work could centre around facilitating improved preparedness on campus in case of an emergency or crisis, and assessing the mental resilience of the student body in response to adverse events that affect the mental well-being of students.

Read the original document here