Social media and teenagers: what the evidence actually supports

There is a short answer about teenagers and social media and a useful one, and they are not the same. What follows is the useful one.

The averages are small and argued over

Large dataset studies have generally found small average associations between total social media use and wellbeing, and researchers disagree sharply about what they mean. Some analyses put the average association in the same range as variables nobody worries about, which has been a genuine and unresolved argument in the field. A small average effect is not the same as no effect for anyone; it is consistent with large effects on a minority.

That distinction is why blanket claims in either direction should be treated carefully.

Sleep is the clearest route

Time on a phone at night displaces sleep, notifications interrupt it, and an unresolved conversation keeps a teenager awake in a way a book does not. The links between insufficient sleep and mood, concentration and irritability are far better established than any direct link between apps and mood.

This makes bedtime the highest-value place to intervene, and it happens to be the easiest to define. It also gives a reason for a rule that a teenager can actually evaluate rather than simply resent.

What they are doing matters more than how long

Passively scrolling other people’s edited lives and actively messaging friends are different activities that get counted identically. Comparison is the mechanism most often implicated, which is why appearance-focused and image-heavy platforms show stronger associations, particularly among girls. Using the same platform to organise a real social life looks quite different in the data from using it as a broadcast feed.

On a normal day, asking what they actually do on it is more informative than asking how long they were on it.

The vulnerable minority is not the average

Recommendation systems respond to engagement, which means a teenager who lingers on distressing content is shown more of it. Content relating to self-harm, disordered eating and extremism is the specific risk that population averages disguise. For a teenager already struggling, the risk profile is different enough that population statistics should not be reassuring.

Platform safety settings and reporting tools exist, are usually buried, and are worth setting up together rather than covertly.

Bans have a cost worth naming

For most teenagers a substantial part of their social life is organised there, so removing it entirely removes something real. Blanket prohibition also tends to produce concealment rather than abstinence, which costs you the information you most need.

Minimum age rules on platforms exist and are widely ignored, and enforcing them early is far easier than reclaiming ground later. Friction, timing and settings generally achieve more than a rule that will be worked around within a fortnight.

Take what is useful and ignore the rest — nobody is marking this.

What is worth doing

Devices charging outside bedrooms overnight, for the whole household, is the single change with the clearest rationale. Explaining the business model plainly, that attention is the product being sold, gives teenagers something to be cynical about, which they enjoy.

Agreeing in advance what they would tell you about, contact from strangers, threats, images, matters more than monitoring for it. A teenager who is distressed, withdrawn or not sleeping needs a doctor rather than a screen-time policy, and the two should not be confused.

The takeaway

Fix the overnight charging, ask what they do rather than how long, and treat distress as a health question.

The thing they remember is rarely the thing you were worrying about.