What Age Should a Child Join Social Media? The 13 Is Not What You Think
Children typically make a first account at 11 to 13 against a platform minimum of 13 — a number that came from a privacy law, not from research. Here is what the evidence does and does not say.
Children typically create a first social media account between ages 11 and 13 against a platform minimum of 13, a threshold that comes from US privacy law rather than child development research, and the evidence on average mental-health effects remains small and contested.
The question has a number attached to it that almost everyone misreads. Thirteen is not the age researchers converged on. It is the age at which a US privacy law stops applying.
Where the 13 comes from
COPPA — the US Children's Online Privacy Protection Act — requires verifiable parental consent to collect data from children under 13. Meeting that requirement is expensive and legally risky, so platforms set their minimum at 13 and stay outside it.
That is the entire origin of the number. It is a compliance boundary that has been quoted as a safety recommendation for two decades.
| Platform minimum age | 13 on almost all major services |
|---|---|
| Verification method | Self-declared date of birth |
| Typical age of first account | 11–13 |
| EU position on profiling ads to minors | Prohibited under the DSA |
| Countries introducing under-16 restrictions | A growing number since 2024 |
So the practical situation is a stated minimum of 13, no verification behind it, and most children arriving a year or two early. Which means, as with the starting school question, you are setting the age yourself whether or not you intended to.
What the evidence actually says
This is where care matters, because the topic attracts confident claims in both directions and the honest position is less satisfying than either.
On average, the measured effect is small. Across the large datasets, the association between social media use and adolescent wellbeing is real and consistently modest — small enough that some well-known analyses found it comparable in size to effects nobody worries about at all.
The subgroup picture is where the serious disagreement is. Effects appear larger for some groups, particularly girls, and the argument among researchers is about how much of that is social media doing something and how much is social media being where existing distress becomes visible. That argument is not settled, and anyone telling you it is has picked a side.
One adjacent finding is settled, in the negative. Video games causing violence has been tested repeatedly for decades and is not supported. It is worth knowing because it is still asserted regularly, and it is a useful calibration for how confident to be about the newer claims.
One finding does replicate. Evening screen use and delayed sleep onset. That is consistent across studies, has a plausible mechanism, and is the one place where a household rule has evidence behind it rather than intuition.
What actually seems to matter more than age
The variables that keep appearing as more useful than the birthday:
- Public account or private, and whether the child knows the difference.
- Following feed or algorithmic feed. These are different products with different failure modes, and the second one is the one that decides what a child sees.
- Whether the phone leaves the bedroom at night. The one evidenced rule.
- Whether the child will tell you when something goes wrong — which is almost entirely a function of what happened the first time they did.
- Whether an adult has looked at the app themselves. Not read about it. Opened it.
A child who is on a private following-based account, whose device charges in the kitchen, and who has told you about something unpleasant once and was not punished for it, is in a better position at eleven than a child with none of those at fourteen.
The honest summary
There is no evidence-based age. There is a compliance number that got mistaken for one, a set of conditions that matter more than the number, and one genuinely well-supported rule about sleep.
If you want a default, most families land somewhere around 13 to 14 for an algorithmic platform and earlier for messaging with known contacts — and the reason to prefer that shape is not a study, it is that messaging with people you know has fewer ways to go wrong than a recommendation engine does.
Say which one you are choosing and why, rather than quoting thirteen as though someone measured it.
See the underlying figures at social media and children statistics, and the control settings that go with them at parental controls statistics.