Spark

High levels of pandemic-induced anxiety, depression observed in social media posts

As we interact more with our digital technologies, can we analyze behaviour on social media for mental health insights?

Researchers say people also expressed a heightened sense of belonging

A woman is shown holding her nose and looking at her phone in front of a window.
New research suggests signs of anxiety, depression, stress, as well as support-seeking expressions have significantly increased during the COVID-19 period as compared to similar data from 2019. (KieferPix / Shutterstock)

Originally published Dec. 4, 2020.

People experienced increased levels of stress, anxiety and depression during the pandemic this spring, according to a study of social media posts published during the first wave of COVID-19 in the United States.

"What we found is that this pandemic has had a really negative impact on people's mental health states," Munmun De Choudhury, an associate professor of interactive computing at Georgia Tech University in Atlanta, Ga., told Spark. "By looking at social media, we are able to get a glimpse into what these challenges are, and how they have been evolving over time." 

Munmun De Choudhury heads the Social Dynamics and Well-Being Lab, where her research focuses on the intersection of machine learning, social computing and mental health. (munmund.net)

De Choudhury and her team at the Social Dynamics and Well-Being lab analyzed more than 60 million social media posts shared on Twitter in the U.S. between March and May 2020. Researchers compared the posts to tweets published in the same three-month period in 2019. They used machine learning algorithms to examine the language of the posts for signs of depression, anxiety and stress, as well as expressions of support. 

The study was published in the Journal of Medical Internet Research in November.

Koustuv Saha, a computer science PhD candidate at Georgia Tech and the lead author of the study, said that the research also revealed many community-building conversations.

"There were also posts where people are raising awareness about COVID-19 and prevention strategies," he said. "People were posting about arranging for donations for PPE, masks, and other supplies for health workers."

Both researchers noted that the rate of posts eventually declined over the course of three months and has since plateaued, which Saha said could be a result of people adapting to "the new normal, the concerns, the uncertainties of the pandemic."

Social media a source of timely data in a crisis

De Choudhury said that mining social media data for mental health insights allows researchers to get real-time data that other research methods, like surveys, can't provide.

"During a crisis, every moment is important, and we want to bring help to people as they're facing those challenges, not three months later," she said.

Saha added that, unlike a response to a survey question, people's social media posts are unprompted. "If people are feeling bad about a particular thing in a particular moment, they will note a social media post in that very moment, which they might forget in a little while, and they will not record in a survey." 

According to a 2020 factbook published by The Canadian Internet Registration Authority, the not-for-profit agency that manages the .ca internet domain, most Canadians (42 per cent) consider social media neutral to their mental health, while 35 per cent report seeing it as beneficial and 16 per cent think of it as harmful.

We want to bring help to people as they're facing those challenges, not three months later.- Munmun De Choudhury, researcher at Georgia Tech University

De Choudhury says another advantage of social platforms is their accessibility.

"Given that a lot of people are already on social media, that could be the mechanism for them to access something that may not be as rigorous and effective as formal health care, but it could be a stepping stone for them to improve access and get the help that they may need," she said. 

Computer science PhD candidate Koustuv Saha focuses on areas of computational social science and social computing. (koustuv.com)

However, the researchers acknowledged that some populations — for example, people aged 65 and older — are not as active on social platforms as the younger demographics, which presents challenges in assessing their mental health states and offering solutions using this methodology.

These limitations are part of the reason why De Choudhury believes an interdisciplinary approach is necessary for future crisis management. 

"We can no longer think in terms of silos of disciplines. Solving big world problems needs scientists from different backgrounds. If you think about how we have tried to respond to the [COVID-19] crisis, whether it's frontline workers or the vaccine development, scientists all across the world of all different backgrounds have been working together. 

"I believe we have to keep that momentum."

Click the 'Listen' button at the top of the page to hear the full conversation with Munmun De Choudhury and Koustuv Saha.