Children's Sleep Recommendations by Age: What the Numbers Are, and What They Aren't

Nine to twelve hours. That is the figure a parent of a nine-year-old will find in the first three search results, and it is a defensible number. What rarely travels with it is its provenance: thirteen people in a room, working through 864 published papers with a structured voting procedure, arriving at a range wide enough to hold a three-hour spread. Children's sleep recommendations by age are real guidance assembled by careful people. They are not measurements, and the difference matters more than it sounds like it should.

The myth: the hours are a threshold

The way the numbers get repeated — in back-to-school explainers, in pediatric handouts, in the app that scores your toddler's night — implies a cliff. Above nine hours, healthy. Below nine hours, accumulating damage. Parents who are otherwise skeptical people absorb this version, then spend September negotiating with a fourth-grader over a boundary they believe was established in a lab.

It wasn't.

What the guideline actually is

The standard citation is Paruthi and colleagues, published in the Journal of Clinical Sleep Medicine in 2016 as a consensus statement of the American Academy of Sleep Medicine, subsequently endorsed by the American Academy of Pediatrics. A 13-member panel screened 864 articles and used a modified RAND appropriateness method — rounds of anonymous rating until the group converged — to settle on ranges by age band.

That is a transparent and legitimate procedure. It is also, by design, a procedure for extracting expert judgment when the underlying evidence cannot settle a question by itself.

And it can't, for a structural reason. Nearly all of the input data is observational, with sleep duration reported by parents or by the children themselves, set against outcomes like attention scores, BMI, depressive symptoms, and injury rates. Associations. No one randomized a cohort of ten-year-olds to eight hours a night for a decade, and no one should.

The most instructive thing in the document is a refusal. For infants under four months, the panel declined to recommend anything at all — the normal variation was too wide to draw a line through.¹

How much sleep does a child actually need?

The answer is a range, not a number, and the ranges are wide on purpose. Here is what the 2016 consensus recommends per 24 hours, naps included where noted:

A seven-month-old sleeping 12 hours and one sleeping 16 are both inside the guidance. That is not sloppiness. Iglowstein and colleagues, in Pediatrics in 2003, followed 493 children in Zurich from infancy through adolescence and published percentile curves for total sleep; at six months the band from the 2nd to the 98th percentile was roughly five hours wide. Galland's 2012 systematic review in Sleep Medicine Reviews, pooling normative data across dozens of observational studies, found the same shape. Within-age variation is nearly as large as between-age variation.

Which means the chart tells you less about your particular child than the confidence of its presentation suggests.

What changes at puberty, in the order it happens

The one place the developmental biology is unambiguous is adolescence, and it is worth walking through as a sequence.

First, somewhere in mid-puberty, the evening rise in melatonin shifts later — the internal clock's signal to begin the night arrives after it used to. Second, the pressure to sleep that builds across a waking day appears to accumulate more slowly in mature adolescents than in prepubertal children; the commonly cited laboratory work here comes from Carskadon's group and used single-digit group sizes, so treat the effect as directionally credible and precisely unknown. Third, evening light in this window produces a stronger phase-delaying push than it does earlier in childhood.

The result is a teenager who is not tired at 9:30 p.m. and is not performing. Sleep onset drifts toward 11. The first bell does not drift. The hours come off the morning end, which is where REM sleep is concentrated — so the loss is not evenly distributed across sleep stages. Then Saturday arrives, the teenager sleeps until 11, and that late Sunday-morning light exposure delays the clock a little further before the week restarts.

This is why the schedule intervention outperforms the lecture. Dunster and colleagues, in Science Advances in 2018, measured Seattle sophomores with wrist actigraphy before and after the district pushed high school start times back by 55 minutes, and found a median gain of about 34 minutes of actual sleep. Two cohorts of roughly 90 students each — small, single-district, and the strongest available real-world test.

Sorting the tactics by how much is behind them

A consistent bedtime routine has the best support of anything in this category: Mindell's randomized work in young children found that a fixed nightly sequence shortened sleep onset and reduced night wakings within about three weeks. The mechanism is conditioning, not chemistry.

The screens-off-one-hour rule is on thinner ground than its ubiquity implies. Evening light does suppress melatonin — Chang and colleagues showed it in PNAS in 2015 — but that study used 12 adults reading a bright e-reader for four hours before bed. Extrapolating from that to a phone at arm's length for twenty minutes is a leap. The hour is a sensible convention. It is not a dose.

An honest rule of thumb

Set the wake time first and count backward, then let the chart's range be your starting hypothesis rather than your target. Watch two signals for two weeks: whether your child falls asleep within roughly 20 minutes of lights out, and whether they surface on their own most weekday mornings. If bedtime is already generous and mornings are still an extraction, the constraint is the schedule, not the child.

Verdicts

• Well-established: the adolescent phase delay is biological, and later school start times measurably increase sleep.
• Well-established: a consistent, predictable bedtime routine improves sleep onset in young children.
• Plausible but thin: the specific boundaries — why 9 hours and not 8.5 for a ten-year-old. These are consensus lines, honestly labeled as such by their authors.
• Folk wisdom: the precise 60-minute screen curfew, and the idea that a child sleeping at the low end of the range is therefore sleep-deprived.

The chart is a hypothesis about your child. The mornings are the data.