The Ferber method is the most argued-about technique in baby sleep and one of the few with a genuine evidence base behind it — though not under that name. In the literature it is graduated extinction, and almost everything worth knowing about it comes from trials that never mention Ferber at all.
What it actually is
You put your baby down awake, leave, and return at progressively longer intervals to reassure them briefly without picking them up or feeding them. Richard Ferber called it progressive waiting; researchers call it graduated extinction. The word extinction is behavioural jargon for removing the response that maintains a behaviour — here, the settling help — not a description of what it is like to do.
What the evidence supports
The American Academy of Sleep Medicine's task force reviewed 52 treatment studies. Across all of them, 94% reported that behavioural interventions were efficacious, with over 80% of treated children showing clinically significant improvement maintained for three to six months. Graduated extinction is one of the specific therapies the review supports.[3] Those findings were turned into formal practice parameters for children from birth to just under five years.[4]
94% / 80%
The harm question, answered properly
This is the part parents actually want, so it is worth being precise rather than reassuring. A randomised trial put 43 infants aged 6 to 16 months into graduated extinction, bedtime fading or a sleep-education control, and measured infant stress directly through morning and afternoon salivary cortisol rather than by asking parents how it felt.[1]
Cortisol showed small-to-moderate declines in the graduated extinction group compared with controls — stress went down, not up. Twelve months later the children were assessed for emotional and behavioural problems, and the mother-child pairs went through the strange situation procedure to measure attachment. There were no significant differences between the groups on either.[1]
A separate trial followed 326 children to age six, five years after a behavioural sleep programme delivered at eight to ten months. It found no evidence of differences between intervention and control families on any outcome measured: child mental health, sleep, psychosocial functioning, chronic stress, child-parent closeness or conflict, attachment, or parental depression and anxiety.[2] The authors' conclusion is worth quoting in both directions — behavioural sleep techniques have no marked long-lasting effects, positive or negative.
Before you start
- Age. Trials in this area recruit from around six months. Below that there is no evidence base and a lot of feeding still to do.
- Safe sleep first. Own clear, flat sleep space, on the back, nothing in the cot. This is non-negotiable and unrelated to method.
- Not while ill, teething badly, or in the week of a house move or nursery start.
- Feeds decided in advance. Know which night feeds you are keeping, so you are not making that call at 2am.
- Both adults agreed. The method fails on inconsistency more than on anything else.
- The day sorted first. If the last stretch before bed is wrong, no night-time method will hold — check wake window stacking before you start.
The intervals
Ferber's book gives a grid of progressive waiting times. The exact numbers are not magic and the trials used their own — what matters is that the intervals lengthen, and that they are the same every time within a night.
| Night | First wait | Second | Third and after |
|---|---|---|---|
| 1 | 3 | 5 | 10 |
| 2 | 5 | 10 | 12 |
| 3 | 10 | 12 | 15 |
| 4 | 12 | 15 | 17 |
| 5 | 15 | 17 | 20 |
| 6 | 17 | 20 | 25 |
| 7 | 20 | 25 | 30 |
How a night runs
Keep the check-ins short and boring. The point of going in is to reassure you both that nothing is wrong, not to settle the baby — the settling is the part they are learning to do.
- 1
Full bedtime routine, then down awake
Drowsy but genuinely awake. If your baby is asleep when they hit the mattress, there is nothing for them to practise.
- 2
Leave, and start the clock
Wait the first interval. Time it on a phone rather than by feel — distress makes three minutes feel like fifteen.
- 3
Go in for under a minute
Lights low, a few words, a hand on the chest if that helps. Do not pick up, do not feed, do not turn it into a conversation.
- 4
Leave again and lengthen the interval
Repeat with the next number on the row. Keep going until your baby is asleep.
- 5
Use the same row for every waking that night
Restart at the first interval after each waking. Move to the next row tomorrow.
- 6
Do the planned night feeds normally
A feed you decided to keep is a feed, not a failure. Do it calmly, then back to the intervals.
What actually happens
Night one is usually the worst and often much worse than parents expect. Nights two and three commonly get better. Night three or four frequently gets worse again before it improves — an extinction burst, which is the behaviour being tested rather than the method failing. Most families who see it through describe a working pattern inside a week to ten days.
Expect regressions after illness, travel and developmental leaps. A rerun usually takes two or three nights rather than a full week, and that is normal rather than a sign the first attempt did not take.
When it is the wrong method
Graduated extinction is not the only supported option, and it is not the best fit for every baby. A trial of 91 infants aged 9 to 18 months compared checking-in — graduated extinction, with gradual separation — against camping-out, where a parent stays in the room. Both improved sleep and held it at six months, but infant separation anxiety moderated which one worked better: infants with high separation anxiety benefited more from camping-out.[5] If your baby is at the anxious end, the chair method is the evidence-based alternative rather than the soft option.
- You cannot tolerate the crying. That is a legitimate reason, not a character flaw, and there are methods that do not require it.
- Your baby has high separation anxiety, where the trial evidence points elsewhere.
- Bedtime is the only problem and night wakings are fine — bedtime fading targets that directly and involves no crying.
- Nothing has improved after two weeks of genuine consistency. That usually means the schedule, not the method.
The honest summary
Graduated extinction works for most families in the short to medium term, does not appear to raise infant stress hormones, and leaves no measurable trace on attachment or behaviour at twelve months or at five years. It is also not the only thing that works, it is unpleasant for several nights, and a good proportion of infant sleep problems improve on their own — in one trial the control group had caught up by four months.[6] Knowing that should make the decision easier in both directions.[7]
