How to Put a Baby to Sleep: the methods with real evidence

What "putting a baby to sleep" actually means at each age

The first thing the sleep research clarifies is that there is no single answer to how to put a baby to sleep — because what works at 3 weeks is not what works at 3 months, and neither is what works at 9 months. Sleep architecture changes dramatically in the first year. So does the baby's ability to self-settle, regulate temperature, and tolerate even brief separations from a caregiver. A method that is gold-standard at 6 months may be inappropriate at 6 weeks. A method that works at 6 weeks may be unnecessary by 6 months.

The American Academy of Pediatrics, in its 2022 evidence-base for safe sleep recommendations, divides infant sleep guidance into two distinct categories: safety recommendations (which apply uniformly from birth to 12 months and are non-negotiable — back sleeping, firm flat surface, no soft bedding, room-sharing without bed-sharing) and settling recommendations, which are explicitly age-dependent and which have a much weaker evidence base because they are harder to study.

Everything that follows is settling guidance. None of it overrides the safety guidance. A swaddle is a settling tool; placing the swaddled baby on their back on a firm surface is the safety rule.

The five methods with real evidence

1. Continuous white noise at moderate volume (50–65 dB). White noise — broadband sound that masks intermittent environmental noise — has more direct sleep evidence than any other settling tool. A foundational 1990 study by Spencer, indexed in Archives of Disease in Childhood, found that 80% of newborns fell asleep within five minutes of white noise exposure, compared with 25% of controls in silence. The mechanism is straightforward: it approximates the in-utero sound environment (which research estimates was roughly 80 dB), and it prevents micro-arousals from incidental sounds.

The two important parameters: volume and distance. The American Academy of Pediatrics recommends white noise devices be placed at least 7 feet (about 2 meters) from the crib and kept at or below 50 dB — roughly the volume of a quiet shower running in the next room. Louder than that, over months, may pose a hearing-protection risk per AAP-cited studies that flagged some commercial sleep machines exceeding 85 dB at one foot. The signal is meant to be ambient masking, not a stimulus.

2. Swaddling under 8 weeks (with a firm cutoff at the first roll). Swaddling — wrapping the baby snugly with arms contained — replicates the boundary pressure of late pregnancy and meaningfully reduces the startle reflex (the Moro reflex) that wakes newborns repeatedly in their first weeks. A systematic review in Pediatrics concluded that swaddled infants showed longer total sleep duration, fewer awakenings, and shorter time-to-sleep compared with unswaddled controls.

The critical safety parameter, restated in the AAP's current safe-sleep guidance: stop swaddling at the first sign of rolling, typically between 8 and 16 weeks. A swaddled baby who rolls onto their stomach cannot easily roll back, and the swaddle becomes a suffocation risk. The transition out of the swaddle is usually managed via transitional sleep sacks (arms-out or one-arm-out designs) that preserve the warmth and slight pressure cues while restoring the freedom to move.

3. Motion soothing — for the early newborn window. Rocking, walking, baby-wearing in a carrier, and yes, the much-maligned car ride: motion settles young infants by activating the vestibular system, which the developing brain interprets as a calming signal. The evidence here is most robust in the first 12 weeks. Research published in Current Biology documented that crying infants carried by a walking adult showed measurable decreases in heart rate within 20 seconds and ceased crying within roughly five minutes — both effects substantially absent when the adult was sitting and holding the baby still.

The mistake parents are often warned about — "creating bad habits" — is largely a myth in the early weeks. Motion-dependent sleep in a 3-week-old does not predict motion-dependent sleep in a 6-month-old. The capacity to self-settle independent of motion develops gradually and naturally, and is associated more with age and neurological maturation than with anything parents do or don't do in the newborn window. By 4 to 6 months, most babies are ready to transition out of motion-dependent sleep — which is where the next method becomes relevant.

4. Drowsy-but-awake transfers (after roughly 4 months). The strongest behavioral sleep intervention with real evidence is the practice of placing the baby in the crib drowsy — eyes heavy, body relaxed, but not yet fully asleep — and allowing them to complete the transition to sleep in the crib. The rationale is that babies who consistently fall asleep in the same context they will wake into (the crib, with the same lighting, sound, and lack of holding) experience their typical overnight micro-arousals as unalarming and resettle more readily. Babies who consistently fall asleep being held and then wake in a crib often interpret the change in context as an alarm and require parental re-creation of the original sleep conditions to return to sleep.

A review in Sleep Medicine Reviews identified drowsy-but-awake transfers as the single behavioral practice most consistently associated with longer overnight sleep stretches and faster resettling after 4 months. Under 4 months, the picture is murkier — newborns' sleep architecture simply doesn't yet support the kind of consolidation this practice trains. Trying to enforce drowsy-but-awake in a 6-week-old usually backfires. Waiting until 12 to 16 weeks usually works.

5. A consistent low-light wind-down routine. The circadian system that regulates sleep-wake timing matures over the first 3 to 4 months of life. By 12 weeks, most infants show measurable melatonin secretion patterns. Light exposure in the hour before bed has a meaningful effect on this maturation: bright overhead light suppresses melatonin and shifts the circadian clock later; warm, dim light supports it. The mechanism is well-documented in the broader chronobiology literature and applies in infancy with particular force, given how new and fragile the system is.

The Wermom editorial archive — see the wermom.com research desk — documents wind-down routines that are consistent in structure rather than complicated in execution. A bath, a feed, a dim-light cuddle, a short song, and the crib transfer, in the same order every night, is more effective than any single elaborate component done occasionally. The repetition becomes the signal. Babies, even young ones, are highly tuned to pattern.

The three methods that don't survive scrutiny

1. Weighted sleep sacks and weighted swaddles. Marketed as inducing "deeper sleep" through gentle compression, weighted sleep products have proliferated commercially in recent years — but the AAP, in its 2022 updated safe-sleep recommendations, explicitly recommends against any weighted product placed on or near an infant during sleep. The concern is mechanical: weight on the chest can subtly restrict respiratory effort in a way that an awake infant easily compensates for but a sleeping infant may not. No randomized trial has demonstrated improved sleep outcomes from weighted products that would justify the safety risk.

2. Essential-oil diffusers in the nursery. The two issues with diffusing essential oils (most commonly lavender, marketed as calming) around infants are documented across the toxicology literature. First, several common diffuser oils — including eucalyptus, peppermint, rosemary, and camphor-containing blends — are explicitly contraindicated in children under 2 due to respiratory and neurological risks. Second, even "safe" oils like lavender lack any controlled-trial evidence of improving infant sleep, and poison-control centers have documented exposures from infant essential-oil contact. The risk-benefit calculation is negative.

3. Rigid wake-window timing in the first 4 months. Wake windows — the amount of awake time between sleeps — have become one of the most popular frameworks in newborn sleep advice, and beyond about 4 months, the framework is reasonable. Under 4 months, the evidence is much thinner. Newborn sleep pressure varies enormously from baby to baby and from day to day, often by 30 to 60 minutes in either direction from any "rule." Enforcing a 60-minute wake window on a 3-week-old who is happily alert at 80 minutes — or anxiously trying to keep awake a 3-week-old who is drowsy at 30 minutes — generally produces more crying and worse sleep than simply following the baby's tired cues. The Wermom editorial team's view, drawn from advisor consensus and reflected in our evidence-first content principle, is that wake windows are a useful framework after 16 weeks and a misleading one before.

The single most useful thing nobody tells new parents

Newborn sleep is biologically chaotic. The first three months are not a sleep-training window. They are a "keep everyone safe and reasonably rested while the circadian system matures" window. Almost everything that "works" in the first three months works because it accommodates rather than overrides the underlying physiology. The interventions with real evidence in the newborn period (white noise, swaddling, motion) all work by mimicking the in-utero environment. They are not training tools. They are bridging tools.

Real behavioral sleep training — drowsy-but-awake, sleep-association management, scheduled night feedings — only becomes appropriate around 4 months, when sleep architecture starts to consolidate and the baby is neurologically capable of the kind of pattern formation behavioral approaches rely on. Before then, the right answer to "how do I put my baby to sleep" is almost always: by responding to the baby in front of you, with the tools the evidence supports, and waiting.

The shorter answer, for the parent reading this at 11 p.m.

Tonight, for a baby under 12 weeks: try the swaddle, the white noise machine 7 feet away at moderate volume, the dimmest light possible, and the gentlest motion that calms — walking, rocking, baby-wearing. Don't try to keep them awake to "fix" their schedule. Don't try to put them down drowsy-but-awake yet. Bridge tonight. The training window comes later.

Tonight, for a baby 4 months or older: try the same wind-down sequence in the same order at the same time. Lower the lights an hour before. Put them down drowsy but not asleep. Resist the urge to immediately re-rescue at the first whimper — give 30 to 60 seconds first, because most micro-arousals self-resolve without intervention. The consistency, more than any single element of it, is what teaches the brain to sleep.

You are not creating bad habits. You are bridging. And the bridge eventually becomes the road.