Sleep Training Methods Compared

Six methods get named in the published literature, three have randomised evidence behind them, one has none at all, and nobody has run the head-to-head trial that would say which is best. This page describes each, what its evidence measured, and what it does not cover. It takes no side, because the published work does not support one: the strongest evidence is not the same thing as the best fit for your household.

What "sleep training" means in the literature

Narrower than the phrase suggests: behavioural interventions for bedtime problems and night wakings — things a parent does at bedtime and in the night, tested against a control group and measured by how fast a child settles and how often she wakes.

The strongest evidence here is a task force appointed by the American Academy of Sleep Medicine, which went through 52 treatment studies. It reported that 94% found the interventions effective, and that over 80% of treated children improved in a way that held at follow-up of three to six months. Over 80% is most children, not all; three to six months is how long they looked, not how long it lasts.

Unmodified extinction

You settle her, put her down awake, leave, and do not go back except for safety — no check-ins, no intervals, until morning. It is the hardest here to sit through, and most families say no.

The task force review gives it strong support, alongside parent education — the two strongest-supported categories in it.

What it does not cover: no trial compared it head to head with another method, so strongest support means better-evidenced rather than better. None of the trials here enrolled infants under six months.

Graduated extinction

You put her down awake and return on a rising schedule of intervals — a short wait, then longer, then longer again — going in briefly without lifting, feeding or starting a conversation. This is what people mean by Ferber, and the Ferber method page covers the name.

The review gives the category further support. Underneath it sits the trial that measured what happened while it ran: infants aged 6 to 16 months randomised to graduated extinction, bedtime fading or a sleep-education control. Both active methods produced significant sleep benefits over the control, cortisol declined modestly in both, and at twelve months there were no differences in attachment measured by the strange situation, or in emotional and behavioural problems.

What it does not cover: nobody tested the famous interval table. Three minutes, then five, then ten across a week — that comes from a trade paperback, and no publisher this site may cite reprints it. The trials tested the category, not the numbers, which is why the interval chart printable here ships them as editable defaults, attribution printed on the paper.

Parental presence, or camping out — the chair method

You stay in the room: sit beside the cot, settle her with as little interaction as the night allows, and over about a fortnight move the chair further away and then out of the door. It takes weeks rather than nights. The chair method page has the steps.

Its evidence is the weakest here, and honestly described it is indirect. The review examined the family this belongs to; the nearest trial taught controlled crying over three consultations rather than a chair. It is worth knowing for its shape: in 156 mothers of infants aged 6 to 12 months, more problems had resolved at two months — 53 of 76 against 36 of 76 — and the difference had gone by four. Depression scores fell further in that group, and for mothers scoring 10 or more on the Edinburgh postnatal depression scale it still held at four months.

What it does not cover: the chair itself. No trial this site can cite has tested staying in the room on its own against anything, so what stands behind it is the family rather than the method.

Pick-up-put-down

You put her down awake; when she cries you lift her, hold her until she calms, and put her down the moment she does. Then again. It is the gentlest-sounding method here and the most physically demanding.

What the evidence covers: nothing. The review's 52 studies do not carry it as a category and no randomised trial has tested it. It is usually offered as the evidence-friendly gentle option, and it is the only one here with no evidence at all.

That is not evidence it fails — nobody has looked — but it means running on reports rather than findings. The pick-up-put-down page covers how families run it and where it breaks down.

Bedtime fading

You move bedtime to the hour she genuinely falls asleep — often much later than the bedtime you have been attempting — so she goes down tired enough to settle fast, then walk that time earlier in small steps. There is little crying to go in for, so there is no decision about going in.

It was the second active arm in the trial above, with the same results: significant sleep benefits over the control, cortisol declining modestly, nothing found at twelve months. The review separately supports bedtime fading with positive routines, so it is evidence-backed as well as quiet. The bedtime fading page has the arithmetic.

What it does not cover: the trial reports both active arms beating the control and does not tell you either beat the other. It needs a late bedtime you can tolerate and a wake-up time you can hold.

Scheduled awakenings

You wake her briefly yourself, a little before the time she reliably wakes, settle her, and stretch the gaps between those planned wakings until they stop. It only works on a waking that arrives at a predictable time, which is why it is rare.

The review gives it further support as a category, and that is the whole of its published backing. How far ahead to go in, and how fast to stretch the gaps, is published by nobody this site may cite; the numbers in circulation are a rule of thumb.

What it does not cover: everything else — no protocol, no age guidance, no measure of what an alarm in your own night costs. A fortnight of notes on a printed day sheet shows whether the waking is predictable enough to consider.

The harm question, in both directions

Two trials have gone looking. The trial of infants aged 6 to 16 months measured salivary cortisol through the intervention: it declined modestly in both active conditions, and at twelve months there were no differences in attachment or in emotional and behavioural problems. A separate trial randomised children whose sleep problems were identified at seven months and followed them to age six: no evidence of differences in mental health, sleep, stress regulation, the child-parent relationship or maternal mental health, and its authors concluded that behavioural sleep techniques have no marked long-lasting effects, positive or negative.

It cuts both ways: nobody has shown lasting harm, and nobody has shown lasting benefit. There is a limit on top of that. Not finding an effect is weaker than proving there is none — those trials looked for particular outcomes, in a limited number of children over a limited span, and something small or rare need not have surfaced. The page on the crying evidence works through it claim by claim.

Choosing, which this page will not do for you

The criteria are your household's: how much crying you can be consistent through, whether two adults agree, whether anybody can hold a plan for a fortnight, and what happens to you at four in the morning if it goes badly. None of that is in any trial, and all of it decides the outcome more reliably than the choice of method.

Two things are worth doing first. A consistent bedtime routine is the one step here with a randomised trial to itself, involves no method and no crying, and is often enough on its own. And check the timing: every trial enrolled babies of six months or older, and when to start covers what that says about a younger baby. If the fortnight ahead holds teeth, a cold or travel, the timeline of the harder stretches will tell you to wait.

Whatever is chosen, the sleep itself does not change: babies go down on their back for every sleep, day and night, on a firm flat mattress, in a cot clear of blankets, pillows, bumpers and soft toys, in your room for at least the first six months. Nothing weighted goes on or over a sleeping baby. Safe sleep basics is the page that never becomes optional.

When this is not a method question

Night waking on its own is not a disorder. The AAP describes a good sleeper at this age as a child who wakes frequently and gets herself back to sleep, rather than one who lies undisturbed for ten hours. Choosing no method at all is a complete answer, and nothing here argues against it.

What is worth a call rather than a plan: a baby who cannot be settled at all, who is feeding poorly or not gaining weight, who is unusually sleepy or hard to rouse, whose cry does not sound like her usual one, or who seems to struggle for breath asleep. A parent who is not coping is a reason to ring somebody today rather than try harder tonight. Speak to your health visitor or pediatrician. This page is general information, not medical advice, and it cannot see your baby.