A researcher from Stellenbosch University discovered that social media can amplify negative emotions, especially among people suffering from depression. For those battling depression, the internet offers a platform to speak openly, find like-minded individuals, and feel less alone—something that is sometimes difficult to achieve in real life.
However, these online connections can also have a dark side: negative feelings can mutually reinforce each other through interactions on social media. This phenomenon was studied by Dr. Kurt Marais, a lecturer in the Department of Logistics at Stellenbosch University, as part of his recently completed doctoral dissertation in operations research.
Marais focused on how emotional states shift within online communities, paying particular attention to depression, and investigated methods for measuring and better understanding this process. His study centered on a subreddit dedicated to depression on the Reddit platform, where users from different countries, age groups, and social strata share their experiences, seek advice, and offer support to others.
Over 56 days, Marais analyzed 12,133 unique posts. The average length of these posts was 178 words, although the longest contained 5,192 words. Most publications came from users in the United States, the United Kingdom, Canada, India, and Germany.
Among the most frequently occurring words were 'like', 'feel', 'life', and 'want', reflecting users' attempts to describe their emotional experiences and desire to feel more than just depressed. There was also frequent use of first-person pronouns such as 'I', 'am', and 'me', while the words 'feeling', 'depression', and 'thinking' appeared frequently alongside 'feel', 'depressed', and 'think', respectively.
The study also revealed patterns in when people most often posted. Activity increased at the beginning of the week, with Mondays and Tuesdays recording more posts than other days, whereas Saturdays were relatively quieter. The day of the week also influenced user reactions: Monday posts attracted more upvotes, while Saturday posts received more comments.
Marais also found that the online community became particularly active late at night. He noted: 'It is interesting that the number of posts increased from 10 PM to 5 AM. This confirms previous research suggesting that people with depression tend to be more active on social media at night.'
One of the most significant findings was the influence of users' emotions on each other. Marais calls this emotional amplification—a process where people can intensify feelings expressed by others in their online network. For example, if two people post negative content, they are likely to continue doing so more frequently and for longer, thereby exacerbating these emotions. This effect was even stronger among people with depression. Although it is also observed with positive emotions, the effect is not as strong as it is with the exchange of negative emotions.
To more accurately define these emotional models, Marais developed a depression-specific lexicon, which is a specialized list of words and phrases related to depressive emotional vocabulary. He found that this tool performed better than general sentiment analysis tools and could identify relevant language in posts about other mental disorders, including anxiety and PTSD.
Marais also compared his approach with large language models trained on mental health materials, arguing that his specialized lexicon offers a more reliable way to detect depressive moods in this context. The study was not limited to Reddit. Using insights gained from the online community, Marais created an agent-based simulation model to study how mental health-related content might spread on Twitter, now known as X. The model was based on observed user behavior, realistic interaction models, natural language processing, and potential intervention strategies. The tweets used in the study were collected before Twitter rebranded to X, so the results pertain to the platform's design, features, and algorithms at that time.
The simulation showed that people's emotional states can persist over time, and interaction with others can also influence these states. Marais discovered that repeated exposure to similar negative emotions can strengthen and deepen these feelings.
According to Marais, the findings have implications beyond academic research. Researchers gain tools to study online mental health; practitioners gain insight into contemporary experiences of depression; platform developers receive evidence-based recommendations to promote user well-being; and social media users themselves better understand their role in algorithm-driven environments. This study underscores the complex role that social media can play in the lives of people experiencing depression. Online platforms can provide connection and support, but the same networks can also create conditions where negative emotional patterns are constantly reinforced. Marais believes his work can contribute to a better understanding of this relationship and help shape a healthier digital environment.