The Surgeon General's 2-Hour Screen Rule, and Its 68 Citations

A parent standing in a kitchen doorway, negotiating the end of one more video, has spent years being told that the research is complicated. On May 20, 2026, the Office of the Surgeon General offered a number instead: two hours a day for anyone between 6 and 18, less than an hour a day for children under 6, and none at all before 18 months.

The document is titled Surgeon General's Warning on the Harms of Screen Use: An Advisory and Toolkit on How to Protect Children and Adolescents, and it arrived with a letter from Health and Human Services Secretary Robert F. Kennedy Jr. Its picture of the status quo is drawn in the same broad strokes as its remedy. Toddlers, it says, average "about two hours using media daily." Pre-teens and teens average "four or more hours per day." The distance between the observed four-plus and the suggested two is the entire policy.

It is worth reading the recommendation's own grammar before repeating it. The advisory does not say children should get two hours. The sentence reads: "Suggested screen time limits could be: none for children under 18 months old, less than 1 hour per day for children under 6, and 2 hours per day for 6–18-year-olds." That is a conditional wrapped around a suggestion, and most of the coverage unwrapped both.

The document defines itself in plain terms: "A Surgeon General's Advisory is a public statement that calls the American people's attention to an urgent public health issue and provides recommendations for how it should be addressed." Nothing in it binds a school district, a pediatrician, an app store, or a household. It contains no language establishing regulatory or legal force, because it has none. An advisory is an instrument of attention.

It is also, by its own account, not an evidence synthesis. The methodology note states that "the findings in this document are not the result of a formal systematic review," describing instead a substantial review of available literature that gave priority to meta-analyses and systematic reviews already published. The reference list runs to 68 entries. That is a reading list assembled to support a position, which is a legitimate thing for an advisory to be, and a different thing from the systematic reviews it draws on.

There is one institutional oddity that most coverage passed over quickly. The office issuing the warning has had no confirmed Surgeon General since January 2025. The advisory arrived under a letter from Kennedy and was written, The New York Times reported, by staffers in the surgeon general's office and the Department of Health and Human Services, including a close Kennedy adviser. The paper ran its story under the headline "A vacant surgeon general's office issues a warning about screen time." The advisory is real and the office is real. The person whose name such documents ordinarily carry is not currently sitting in the chair.

Buried in the text are the sentences a careful reader should hold onto. "Most available evidence is observational and findings vary depending on the type, content, context, and duration of use." "The field is also broad and uses inconsistent methodologies, and research on newer digital media environments is still evolving." On the specific question of screen exposure and social communication differences, it concedes that "current evidence does not establish a causal relationship."

That last admission is not a technicality, because the most plausible alternative explanation runs in the opposite direction from the advisory's framing. Candice Odgers, a psychology professor at the University of California, Irvine who studies the health effects of screen use, has argued that much of the existing research does not suggest social media use is a major predictor of whether a child develops mental health problems; it suggests instead that children already struggling are more likely to spend time online. John Torous, who directs digital psychiatry at Beth Israel Deaconess Medical Center, said the report may read as more certain than the underlying evidence warrants.

The numbers the advisory does supply are also doing four different jobs at once. Nearly five in ten teenagers have experienced cyberbullying — a prevalence figure. Between 3.5 and 5 percent of children experience sextortion before adulthood — a lifetime risk estimate. Nearly 40 percent of children are projected to be myopic by 2050 — a demographic forecast about eyes, not minds. Ninth and tenth graders who use social media frequently are less likely to report mostly A and B grades than low-use peers — a cross-sectional association with an obvious two-way arrow. None of them, individually or together, yields a two-hour threshold. The two-hour figure has no study attached to it in the document.

The first ask is the phone ban: no personal phones for the length of the school day, passing periods and lunch included. Here the advisory hedges precisely, and correctly: "In older children, bell-to-bell phone restrictions may reduce classroom distraction and improve academic outcomes for some students." Two references support it. It is also, at this point, less a proposal than a ratification — at least 37 states and Washington, D.C. already restrict phone use during the school day, 28 of them bell to bell.

The second ask is much larger and rests on much less. Schools are urged to invest in physical textbooks, prioritize pen-and-paper curricula, assign work "in books or on paper whenever possible," and store digital devices in computer labs rather than issuing them to individual students, with exceptions for students whose education plans require digital tools. Evidence that adolescents are distracted by their own phones does not transfer to a claim that district-issued laptops harm learning; those are different devices, different uses, different studies. Education Week notes the recommendation also sits awkwardly beside the same administration's enthusiasm for artificial intelligence in classrooms, and that educators surveyed by its research center split — majorities report negative effects on social-emotional skills and behavior, while more than half also report that digital tools improved engagement and learning. Iowa, Tennessee and Utah passed laws in 2026 pushing schools toward non-digital tools.

The most consequential passages may be the ones aimed at neither parents nor principals. The advisory asks technology companies to eliminate autoplay, infinite scroll and algorithmic recommendations for children, disable notifications on minors' accounts, block direct messages from strangers, default privacy settings to their strongest, and "display a warning about harmful screen use before each use." That is where advisories have historically done their work: not by changing what happens at dinner, but by giving legislatures and product teams a federal sentence to quote. Read that way, the two-hour number is less a finding than a default — a starting position for a negotiation that previously had none.

There is a category of official document that is more useful than it is rigorous, and this is a clean specimen of one. The advisory is honest about itself in ways its coverage was not: it says it is not a systematic review, it says the evidence is mostly observational, it says causality is unestablished, and it phrases its famous number as a conditional. Then the headlines converted "could be" into "should be" and a suggestion into a rule. The school recommendation splits in two, and readers should split it the same way — bell-to-bell phone bans have some support and are already law in 28 states, while the call to retire classroom laptops for textbooks is a policy preference riding on evidence collected about something else. What the advisory genuinely delivers is a shared default where none existed, and a set of design demands aimed at platforms that no household could win on its own. That is worth something. It is not the same as knowing that two hours is the right number, and nobody currently does.