We need your help to keep this website free. You can 
A new Gallup analysis finds no simple worldwide relationship between time on social media and wellbeing. The pattern varies with country income and with the outcome being measured. Heavier use is associated with more negative emotions, but that is not the same as proving that an app caused them, or that everyone faces the same risk.
Source: Ed Gregory / Pixel.la, via Wikimedia Commons, CC0. Illustrative photograph of a phone and laptop, cropped; the blank screen is not a Gallup survey or an app screenshot.
What did Gallup actually measure?
The new analysis uses adult survey data from 97 countries. Its comparisons distinguish low use, including no use or less than 30 minutes a day, moderate use from 30 minutes to under two hours, and high use of at least two hours. Researchers adjusted for factors including age, income, education and social support.
High users reported more negative emotions than moderate users. Yet their evaluations of their lives were not meaningfully worse in the overall global comparison. The lower life-evaluation association appeared in high-income countries, rather than following one uniform pattern everywhere.
Those are separate findings. Feeling stress or loneliness and rating one's life as a whole are different measures; combining them into a single claim that social media makes people unhappy would lose an important distinction in the results.
Does two hours become a personal safety limit?
No. Two hours is a boundary used to group survey answers, not a clinical recommendation. Gallup's report materials describe an observational study. They do not establish whether use changes wellbeing, wellbeing changes use, or other circumstances influence both. This is the familiar problem of correlation versus causation.
The data also do not separate platforms or describe what people were doing online. Messaging a friend, watching videos and scrolling through a feed can occupy the same amount of time without being the same activity. Gallup did not directly measure those experiences here, so the report cannot rank apps by their effects.
What can users take from the findings?
The useful starting point is the scope of the evidence, not an app blacklist. The report's early 2026 dataset covers 97 countries; its materials say fieldwork is expanding toward 140. Readers should not confuse that developing social-media dataset with the wider country coverage of earlier emotions reports.
Our coverage of Pew's survey on social media and democracy concerned public attitudes, a different question from this wellbeing analysis. Likewise, the connections discussed in our article on meeting people through social media cannot be reduced to minutes on a timer. The new findings make a stronger case for asking what was measured than for declaring one verdict on social media.
