Bedtime Fading

Bedtime fading starts the night at the hour she actually falls asleep rather than the hour you would like her to, and then walks that time earlier by fifteen minutes at a stretch once she is going off quickly. Nothing else changes: same routine, same room, same cot. It is one method among several, it was one of the two active arms in a randomised trial of infants aged 6 to 16 months, and the whole of it is arithmetic and patience rather than a decision about whether to go back in.

It is also the option almost nobody has heard of, which is why this page spends its space on how to run it rather than on arguing for it. For the other five side by side, the methods compared page does that without picking one.

Why the late bedtime is the trick, not the problem

Most families arrive at this method with the same evening: she goes in at seven, and at twenty past eight she is still awake, cross, or being resettled for the fourth time. The instinct is that she is fighting bedtime. The likelier reading is that seven o'clock is not yet a time she can fall asleep at, and an hour of failing to has taught both of you that the cot is where nothing happens.

Fading starts from the opposite end. You put her down when the drive to sleep is already high — the time she has actually been dropping off, night after night — so that settling takes minutes instead of an hour. Once that is what going to bed means, you move it. The reason this is dull to run is the reason it works: you are never asking her to do something she cannot yet do.

The arithmetic behind it is worth seeing. The sleep panel whose figures this site uses puts infants of 4 to 12 months at 12 to 16 hours per 24 including naps, and gives no figure at all under four months because the variation is too wide to be useful. Subtract what she sleeps in the day and the night has a length, whether or not the clock agrees with it. A bedtime set two hours in front of that length is not a bedtime; it is two hours of lying down awake.

Finding the starting bedtime from your own notes

This is the part that cannot be skipped, and it is the reason the method takes a week before it starts. You need the time she falls asleep, not the time she goes into the cot, for four or five nights running. Those are usually different by half an hour and occasionally by ninety minutes, and nobody remembers accurately in the morning.

Write it down as it happens, on a phone or on a printed night sheet. Then take the latest of the four or five, not the average — starting from the average means starting too early on the bad nights, and the bad nights decide whether this works.

The notes are worth more than the number they produce, because beside each bedtime sits the length of the last awake stretch — and that is the figure that quietly ruins this method. The wake windows by age guide explains where those bands come from and why they are bands rather than numbers, and the wake window calculator puts them on your clock.

Keep everything else identical while you do it. Same short routine in the same order, lights down, room dark, and into the cot awake — the bedtime routine page covers the one part of this subject that was tested entirely on its own and worked.

Walking it earlier, and how fast

The rule families use is fifteen minutes after three or four good nights, and a good night means she fell asleep quickly at the current time. Three or four rather than one, because a single quick night proves nothing, and fifteen rather than thirty because a half-hour jump is often just far enough to be beyond her.

When a move does not take — she is awake in the cot for half an hour again — put the bedtime back to where it was working and hold it for another three or four nights before trying the same step a second time. Going backwards a step is the method running correctly, not a setback.

How long the walk takes depends on how far you started from where you want to be, and nobody has published a figure for it. A fortnight to a month for a two-hour shift is what families describe; that is a convention rather than a finding.

The response-cost variant, named for what it is

You will see this written as bedtime fading with response cost. It adds one rule: if she has not fallen asleep within about twenty minutes, she comes out of the cot for a short while — awake, quiet, boring, no play — and then goes back in. The reasoning is that lying awake in the cot is exactly the habit the method exists to avoid, so the cot is removed while she is not using it.

Say plainly what that is. The category the task force review supports by name is bedtime fading with positive routines, and the trial arm described below was bedtime fading; response cost is a variant nobody cited here tested separately. Plenty of families never need it, because a starting bedtime chosen correctly rarely leaves twenty minutes of awake time to manage.

What the evidence covers, and what it does not

Two things stand behind this method and both have edges worth knowing.

The first is a randomised trial of infants aged 6 to 16 months, in which bedtime fading and graduated extinction were compared against a sleep-education control group. Both active methods produced significant sleep benefits over the control; salivary cortisol declined modestly in both; and at the twelve-month follow-up there were no differences in attachment measured by the strange situation, or in the children's emotional and behavioural problems. The second is the American Academy of Sleep Medicine task force that went through 52 treatment studies, which reported that 94% found the behavioural interventions efficacious and over 80% of treated children improved in a way that lasted 3 to 6 months, with support for bedtime fading with positive routines as a category.

Now the gaps, because they matter more here than the findings. Nobody has run bedtime fading head to head against the other methods, so anyone who tells you it is slower or gentler or more durable than intervals is estimating. The trial enrolled infants of six months and over, so it says nothing about a four-month-old. It did not measure how many families found it too slow to keep going. And the follow-up ran a year, which is a year rather than a childhood — reassurance about what was looked for rather than proof about everything.

The two ways it goes wrong

The first is starting too early, which is the same mistake the method was designed to remove, made a second time. If the first bedtime is set at the time you wish she slept rather than the time she does, you get a settling fight on night one and the whole advantage is gone. The tell is simple: more than about twenty minutes to fall asleep on two nights running means the starting time is wrong, not that she is.

The second is the day rather than the night. If the last nap ends at four and the faded bedtime lands at half past nine, that awake stretch is longer than she can carry, and a baby who arrives at bedtime past the point of settling looks exactly like a baby who is not tired enough. That is why the naps come first: cap the gap by moving the last nap later or adding a short one, then fade the bedtime. The sleep schedule generator draws the whole day so you can see where the gap has opened, and it prints bands rather than a timetable.

Whatever the timing does, 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. The safe sleep basics page has the full version.

Who this does not suit, and when it is not a bedtime problem

It suits least the household that needs the change this week: it is the slowest option here, measured in weeks rather than nights, and a family under real pressure may find that too costly. It suits least the baby whose bedtime is already early and quick — if she falls asleep in ten minutes and the problem is at two in the morning, there is nothing to fade. And it is a poor fit for a fortnight with a holiday or a house move in it, because the method rests on the bedtime staying where you put it.

A baby who cannot be settled at all, who is unusually sleepy, feeding poorly, not gaining weight, or whose cry does not sound like her usual one is not a timing question, and neither is a worry that will not go away. That is a conversation with your pediatrician or health visitor. This page is general information, not medical advice, and it cannot see your baby.