Buchanan, K., Aknin, L. B., Lotun, S., & Sandstrom, G. M. (2021). Brief exposure to social media during the COVID-19 pandemic: Doom-scrolling has negative emotional consequences, but kindness-scrolling does not. PloS one, 16(10), e0257728. https://doi.org/10.1371/journal.pone.0257728

About the study:

• Before this study, research showed that there was a link between mental health and social media use—especially during the pandemic. It wasn’t clear whether people with poorer mental health sought out more negative news or whether the negative news worsened mental health. (This is what is meant by “correlation does not equal causation”—we knew there was a link, but no studies were yet designed to sort out if one thing *caused* the other.) This is why we put more weight on randomized control trials (RCTs) than we do just about all other types of studies. Well-designed RCTs can help sort out if one thing causes another. This study on doom-scrolling was designed to find out by randomly assigning people to groups and having them watch curated feeds of social media and then using standardized tests to measure their emotional state.
  
• Here’s a blog post I wrote on the concept of “correlation does not equal causation”: https://leesaklich.com/health-research/correlation-does-not-equal-causation/
• Previous studies on the subject of doom-scrolling asked people how much time they spent on social media (they didn’t measure it). We know that people are naturally not great at accurately estimating things they did in the past, like “time on social media” or “how much broccoli you ate in the past 3 months.” This trial was designed to have a timer on the social media exposure so that all participants watched their social stream (depending on which group they were in) for 2-4 minutes. This short timeframe also allowed researchers to learn about the effects of even brief exposures to different types of posts on social media.
• “We examine the emotional consequences of exposure to brief snippets of COVID-related news via a Twitter feed (Study 1), or a YouTube reaction video (Study 2). Compared to a no-information exposure group, consumption of just 2–4 minutes of COVID-related news led to immediate and significant reductions in positive affect (Studies 1 and 2) and optimism (Study 2). Exposure to COVID-related kind acts did not have the same negative consequences, suggesting that not all social media exposure is detrimental for well-being.”
• Conclusion: “Although information-seeking is generally an adaptive coping strategy in times of threat, doing so during a pandemic may be less helpful. Unlike most world events, the threat of the current pandemic affects many life domains (relationships, education, work, leisure), and there is uncertainty about how long it will last, and what will happen next. Even a few minutes of exposure to COVID-related news on social media can ruin a person’s mood. We would all do well to be mindful of these effects and consider balancing our doom-scrolling with some kindness-scrolling.”

• Because this was an RCT, this study had several strengths including the above-noted designs that people were randomly assigned to a group, their positive and negative affects and optimism were measured right afterward, and the exposure was limited to 2-4 minutes.
• While this study didn’t, other studies have shown elevated moods with positive news. More research is always recommended to dive even deeper into these issues and try to answer these questions under different conditions (e.g., 10 minutes of exposure, or platforms other than Twitter and YouTube, or different measures of mental health, etc.) to see if the mental health effects all still point in the same direction or not. 😀

• Therefore, we can say that, according to this study, just 2-4 minutes of COVID-19 news on social media in the midst of a pandemic can decrease your mood. If more, bigger, or stronger studies test for this in different ways and corroborate this finding, we will have a much more robust evidence base to make even more definitive statements.
• Study strength is rated a 6/7 according to this chart (randomized control trial): https://www.compoundchem.com/2015/04/09/scientific-evidence/