How to sleep train a baby: a calm, evidence-based guide

Sleep training has a reputation it does not quite deserve. It is not one harsh technique, and it is not something you either do or refuse to do. It is a small set of well-studied ways to help a baby learn to fall asleep on their own — and the part that does most of the work is the least dramatic bit: a short, predictable bedtime routine, done the same way every night, by whichever parent is on duty.

This is a plain guide to what sleep training means, when it is appropriate to start, how the main methods differ, and what the research actually shows. It is educational, not medical advice — talk to your pediatrician before you begin, especially if your baby was premature, has reflux, or is not gaining weight as expected.

What sleep training actually is

Falling asleep is a skill. Adults have it so thoroughly that we forget it was ever learned: we notice we are tired, we lie down, and at some point we are gone. A baby who has only ever fallen asleep while being fed, rocked, or held has learned a different version of that skill — one that requires another person to be present. When they surface between sleep cycles in the night, as everyone does, they need that person again to get back down.

Sleep training is simply teaching the other version: that the crib, in the dark, at a predictable time, is a place where sleep happens without a parent doing the settling. That is the whole of it. The different "methods" are just different speeds at which you step back.

What it is not: it is not about night weaning, which is a separate decision on its own timeline. It is not about leaving a hungry or unwell baby. And it is not a permanent state — every family in the world regresses through illness, teeth, travel, and developmental leaps, and every family finds their way back.

Before you start: safe sleep comes first

No method matters if the sleep space is not safe. These rules apply from the very first night, and they are the non-negotiable foundation under everything else:

• Back to sleep, every sleep, until age one — including preterm babies and babies with reflux.
• A firm, flat, level surface — a crib, bassinet, or play yard. Never an incline, and never a soft or hammock-like surface.
• Bare is best. No pillows, blankets, bumpers, soft toys, pods, nests, or weighted swaddles in the sleep space.
• Room-share, don't bed-share. Baby in your room on their own surface for at least the first six months.
• Don't overheat. Keep the head uncovered and use a sleep sack rather than loose bedding; the NHS suggests a room around 16–20°C.
• Stop swaddling at the first sign of rolling, and don't use car seats or swings for routine sleep.

These come from AAP safe-sleep guidance and the NIH's Safe to Sleep programme, and they are worth re-reading each time your baby hits a new stage.

When to start: the 4-month line

Newborns cannot be sleep-trained, and trying is a recipe for misery on both sides. In the first three or four months, sleep is governed by hunger and a body clock that is still assembling itself. The work of that phase is different: feed on demand day and night, keep the days bright and social and the night feeds dim and boring, follow wake windows of roughly 45 to 90 minutes to head off overtiredness, and put your baby down drowsy-but-awake when you can, so that falling asleep in the crib starts to feel ordinary.

Around four months, most babies develop the maturity to start learning independent sleep, and that is when formal methods become appropriate — with your pediatrician's okay. From there the useful framing is age bands rather than a single plan:

• 0–3 months — settle in. Rhythm, not training. Safe sleep, day-night distinction, and survival.
• 4–6 months — foundations. The classic window to begin. A consistent bedtime routine plus one chosen method.
• 6–12 months — consolidate. Night sleep lengthens and naps settle into a predictable shape.
• 12–18 months — toddler. The two-to-one nap transition, and new independence to work with.
• 18 months–3 years — big feelings, big boundaries. Crib-to-bed, testing, and calm consistency.
• 3–4 years — independent sleeper. Bedtime stalling and getting out of bed, handled with structure rather than negotiation.

The methods, plainly

Every named approach is a variation on how quickly you withdraw your presence. Here are the ones with real evidence behind them:

• Bedtime fading (4+ months). Temporarily set bedtime to the time your baby naturally falls asleep, then move it 10–15 minutes earlier every few nights until you reach your target. It works with your baby's sleep pressure rather than against it, and involves remarkably little crying.
• Graduated extinction, also called Ferber or check-and-console (4–6+ months). Put your baby down awake; if they fuss, wait a set interval before a brief, calm check without picking up, lengthening the interval each time. The best-studied method, and usually effective within three to seven nights.
• Pick-up / put-down (4–7 months). Pick up to calm, then put down awake as soon as they settle, repeating as needed. Low-cry but labour-intensive, and it gets unwieldy once a baby is heavy and mobile.
• Chair method, or gradual retreat (6+ months). Sit beside the crib until your baby sleeps, then move the chair a little farther away every few nights until you are out of the room. Gentler and slower — think weeks rather than nights.
• Full extinction (4–6+ months). After a loving routine, put your baby down awake and do not return until morning, aside from scheduled feeds and safety checks. The fastest results, and the research shows no long-term harm — but only worth choosing if it genuinely fits your family, because a method you abandon halfway is worse than one you never started.
• Silent return (2+ years). When your child leaves the bed or calls out, calmly walk them back with no talking and no eye contact, every single time. Pair it with a sticker chart for staying put.
• Bedtime pass (3+ years, works from about 2.5). Give your child one physical "pass" they can trade for a single sanctioned get-up or request after lights-out; ignore any further bids, and let an unused pass earn a small reward in the morning. It has the strongest evidence for this age, delivering extinction-level results with far less crying.

Notice what none of these lists contain: a claim that one is correct. Pick the one you can actually sustain for a fortnight at 2am on a Tuesday, and give it that fortnight. Switching methods every few nights is the single most reliable way to prolong the crying, because your baby is learning the rule and you keep changing it.

What the research actually says

The worry most parents have is not whether sleep training works — it plainly does — but whether it costs something invisible. That question has been studied directly.

A 2016 randomised controlled trial in Pediatrics (Gradisar and colleagues) compared graduated extinction, bedtime fading, and a control group. Both methods improved sleep. Neither produced any measurable increase in infant stress hormones, and neither affected parent-child attachment. A 5-year follow-up of an earlier trial (Price and colleagues, 2012, also in Pediatrics) tracked children into school age and found no lasting differences in emotions, behaviour, or the parent-child relationship. A 2006 review by the American Academy of Sleep Medicine (Mindell and colleagues) found behavioural sleep interventions effective across a large body of studies.

And then there is the finding that changes how you should spend your effort. In a study of 10,085 mothers, Mindell and colleagues (2015) found a clear dose-response relationship for bedtime routines: the more nights per week a family kept a consistent one, the faster children fell asleep, the fewer night wakings they had, and the longer they slept. More nights, better sleep — a straight line.

Which means the routine is not the warm-up act before the real intervention. In a lot of homes, it is the intervention.

The routine is the part you can control

A good bedtime routine is short, in the same order, and boring in the best possible way. Twenty to forty minutes, three or four steps, ending in the crib or bed. Bath, pyjamas and sleep sack, a feed, a book, a song, lights out. The specific steps matter far less than the fact that they are the same steps, in the same order, every night.

The place this usually breaks is not knowledge. It is handoff. One parent has the routine in their head, complete with the small undocumented details — the second verse of the song, the nightlight at the lower setting, the door left open exactly this far. The other parent does their honest best and gets a slightly different version. From the baby's side, that is two different bedtimes, and the consistency the research keeps pointing to quietly evaporates on the nights the usual parent is out.

Getting the routine out of one head and into a place both parents can follow identically is not administrative fussiness. It is the mechanism.

How Kinrows' sleep training routine builder works

This is exactly what we built the Routines section of Kinrows to solve — and it is the part of the app parents of small children reach for first.

Start a sleep training routine for your child, enter their birthdate, and Kinrows drops you into the right phase automatically. A four-month-old opens on Foundations with bedtime fading recommended and graduated extinction and pick-up/put-down offered as alternatives. A three-year-old opens on Independent sleeper with the bedtime pass. As your child grows, the guidance moves with them — you are never reading advice for a stage you left behind two months ago.

Each phase gives you the same four things:

• A recommended method, with alternatives. Named, explained in a sentence or two, with the age range it suits — so you are choosing between real options rather than guessing from a forum thread at midnight.
• Concrete steps for this phase. Not principles — the actual things to do this week, from wake windows to how to handle the 3am wake.
• The safe-sleep rules, always visible. Surfaced on every phase, not buried in an intro you read once in the newborn fog.
• The sources. Every phase is built against published guidance and trials — AAP and NIH safe-sleep policy, the AASM sleep-duration consensus, and the Gradisar, Price, and Mindell studies above — and the citations are listed in the app so you can read them yourself.

Around that sits the routine builder. You write your bedtime routine once — the steps, the order, the timing — and it lives in the household, not in one person's memory. Whoever is on bedtime duty opens the same routine and runs the same sequence. A grandparent covering for the evening gets the real version, second verse and nightlight setting included. That is the consistency the dose-response research rewards, made reachable on a night when you are running on four hours of sleep.

Alongside it, baby sleep tracking records naps and night stretches so you can see the shape of a week rather than relying on the memory of whoever was awake for it — useful for spotting that bedtime has drifted twenty minutes late, and genuinely useful at a pediatrician's appointment. And because Kinrows is a household app rather than a solo one, both parents see the same picture. Nobody has to be briefed.

The whole thing is educational scaffolding, deliberately so. Every phase leads with "check with your pediatrician", and formal training stays gated until around four months, exactly as the evidence and the AAP recommend. It is free, and it is private: your household's data is isolated, never sold, and never used to train AI models.

When it isn't working

Two weeks of genuine consistency with no movement at all is a signal, not a failure. Before you conclude the method is wrong, check the boring things first: is bedtime drifting later on some nights, is the room bright at 5am, has a nap quietly dropped, is your child overtired by the time you start? Then check the calendar — teething, illness, a house move, a new sibling, and the 8-, 12-, and 18-month developmental leaps all buy you a temporary wobble that resolves on its own.

And if sleep is genuinely not improving, or you see snoring, long pauses in breathing, or daytime symptoms that worry you, that is a conversation with your pediatrician rather than a different method.

Frequently asked questions

What age can you start sleep training?
Around four months for most babies, once the body clock has matured and night feeds are less hunger-driven. Before that, focus on safe sleep and day-night rhythm. Check with your pediatrician first, especially if your baby was premature or has reflux.

Is sleep training safe for babies?
The evidence is reassuring. A 2016 randomised trial in Pediatrics found no effect of graduated extinction or bedtime fading on infant stress hormones or attachment, and a 5-year follow-up of an earlier trial found no lasting differences in emotions, behaviour, or the parent-child relationship. Safe-sleep practices matter far more to your baby's safety than which method you choose.

Which sleep training method works best?
There is no single winner. Graduated extinction and bedtime fading have the strongest infant evidence; the chair method and pick-up/put-down are gentler and slower; the bedtime pass is best for toddlers and preschoolers. The one that works is the one you can apply consistently for a full one to two weeks.

How long does sleep training take?
Extinction-based methods often show clear improvement in three to seven nights, with most families settled inside two weeks. Gradual methods take two to four weeks. Expect an extinction burst — a night or two that gets worse before it gets better.

Does a bedtime routine actually help?
Yes, and the more nights a week you keep it, the better it works. A study of 10,085 mothers found a clear dose-response effect on how quickly children fell asleep, how often they woke, and how long they slept. Consistency, not the specific steps, is what carries the benefit.