Research

How social media affects mental health after COVID

8 min read · Educational only. Not medical advice.

The pandemic did not invent the feed. It handed the apps a captive audience — and changed what people expected from a scroll. A lot of that did not leave when the restrictions did.

What changed

Before March 2020, social media was already a large part of the day. Then it became, for a while, the main way to connect, get news, work, pass time, and leave the room without leaving the room.

DataReportal’s 2021 overview recorded a rise in daily social media use in that first pandemic year — on the order of twenty percent globally. Quantity mattered. Context mattered more. People who had used the apps alongside a full offline life were suddenly using them as a stand-in for one. That is a different relationship.

Adults now average about 4.8 hours of screen time a day (DataReportal, 2024). Common Sense Media’s 2021 census of tweens and teens documented a large jump in media use in the same window. The point is not a single shocking number. The point is that the day reorganized around a rectangle, and then stayed that way.

The isolation paradox

One finding that kept showing up: more social media during lockdown often traveled with more loneliness, not less. Cauberghe and colleagues looked at how adolescents used the apps to cope with loneliness and anxiety in 2020. The apps were a tool. They were also not the same thing as company.

Scrolling other people’s highlights is a passive act. It can look like connection without delivering much of what a real exchange delivers — being seen, answering, sharing a room. Sherry Turkle has been writing for years about conversation thinned by a device on the table. COVID made that the default social life for a long stretch.

A window is not the same as a door. The feed showed other people’s lives. It did not put you in the room.

What the mental-health data showed

This is educational, not a diagnosis, and the picture is mixed in the way real research is mixed. A few things are solid enough to say plainly:

• Twenge and Joiner, using U.S. Census measures, documented a rise in anxiety and depressive symptoms between 2019 and 2020.
• CDC Youth Risk Behavior Survey summaries from the following years described a hard stretch for adolescent mental health. Girls in particular showed up in the more serious numbers.
• Sleep took a hit in many households, and late-night feeds were one of the usual companions.
• “Doomscrolling” — staying in a bad-news loop because the next item might explain the last one — became a word people needed.

A 2022 paper in the American Economic Review by Braghieri, Levy, and Makarin, using the Facebook rollout, added to a longer argument that social media is not neutral for mental health. Correlation is not destiny. It is also not nothing.

The algorithms had a captive classroom

Lockdown gave the platforms something they had not had at that scale: billions of people with nowhere else to go, producing data all day. Recommendation systems got a long, clean training run.

What kept people scrolling was not always what they would have said they wanted. Outrage, comparison, and fear are sticky. The systems learned that. Those systems are still running. The lockdowns ended. The engineering did not roll back.

Patterns that did not unlock

COVID did not invent this. It locked it in. Tracking since then keeps finding versions of the same three things:

• Daily use, for many age groups, did not return to the old baseline.
• Checking during meals, first thing in the morning, and last thing at night — times that used to be empty of a feed — stayed.
• A lot of people report that ordinary social life feels harder than it did. Less practice is part of that. A brain trained on a hotter stimulus is the other part.

A quieter effect: the day feels underpowered

After enough time in a curated, high-contrast feed, a quiet meal or a walk can feel insufficient. That is not a personality collapse. It is a comparison. The feed set the volume. Post-COVID, after years of a louder feed, the effect is easier to notice.

If you want the mechanism in more detail, the dopamine piece is the companion to this one.

First steps that are actually small

Knowing the mechanism puts a little air between you and the mood the feed leaves behind. Then change the defaults:

• Look at the real number. Screen Time or Digital Wellbeing. Estimate, then compare.
• Unfollow hard. If an account consistently leaves you worse, it does not get a seat. The algorithm adjusts.
• Put it on a clock. Two short, chosen sessions are a different habit than an all-day drip.
• Fill the gap. The usual mistake is deleting an app and leaving a hole. The hole wants stimulation, connection, or distraction. Give it a walk, a person, or a page.

The 7-day guide walks those changes one day at a time. If you want a longer stretch later, the 30-day system is $47. Neither one is treatment.

The point

If your sleep, focus, anxiety, or sense of company shifted after 2020, the feed is a reasonable place to look. That is not a personal verdict. It is a system that got a once-in-a-century training period, and then kept the gains.

You do not need a sermon about quitting the internet. You need a few different defaults, starting with the first hour and the last.

All five pieces