The Ferber Method: How It Works and How It Compares

The Ferber method is graduated extinction: you put your baby down awake, leave, and return at gradually lengthening intervals to reassure without picking them up. Dr. Richard Ferber called it progressive waiting.

It is also probably the most misrepresented technique in parenting. Ferber himself says so. Both the people who champion it and the people who condemn it are frequently arguing about a protocol he did not write.

So here is the actual method, the actual check-in schedule, what the research does and does not show, and an honest comparison against the alternatives.

What the Ferber Method Actually Is

Dr. Richard Ferber founded and formerly directed the Center for Pediatric Sleep Disorders at Boston Children’s Hospital. His book Solve Your Child’s Sleep Problems was published in 1985 and substantially revised in 2006.

The mechanism targets what clinicians call a sleep-onset association. A child who is always rocked or fed to sleep learns that sleep onset requires a caregiver, and so signals for one at each normal night arousal. Everyone wakes briefly between sleep cycles. The question is what has to be true for the child to go back down.

The Real Check-In Schedule

Ferber publishes a sample schedule. Note that he presents it as a sample, and says families may adapt it.¹

The clock starts when you leave the room. The same schedule applies to night wakings on that night.

A check-in is brief. Ferber suggests one to two minutes, and the research literature typically describes checking procedures of 15 seconds to a minute. You comfort by voice and light touch. The check-in exists to reassure you and confirm your child is safe. It is not there to stop the crying, and it usually will not.

What age this is for

The commonly stated floor is 4 to 6 months. Be aware of the distinction: 4 months is roughly when sleep architecture and circadian rhythm have consolidated enough for the approach to make sense, but essentially every randomized trial enrolled infants of 6 months or older.²³ Below 6 months there is no trial evidence for extinction-based methods at all.

How the Popular Version Misrepresents Ferber

Ferber has been unusually direct about this. In a 2006 NPR interview he said the concept that some people are implying, that he has an approach he uses for children who are not sleeping well, is a complete misunderstanding.⁴

The method is a targeted treatment for a defined problem, reserved for families who want to change what he calls established sleep associations. It was never proposed as a universal parenting program.⁴

The 2006 edition also reversed his position on co-sleeping. He had previously suggested parents who wanted to co-sleep should examine their own motivations. He now says it works just fine for many families, adding that children are very flexible and can sleep well in many different settings.⁴

He also dislikes the verb “Ferberize” and would rather it went away.

This cuts both ways in the public argument. Critics often invoke a shutdown response in babies left to cry for long periods week after week, which is not something Ferber advocates.⁴ Both sides are frequently arguing about something else.

That said, do not oversell the gentleness either. As one pediatrician reviewer puts it, the Ferber method is essentially the cry-it-out method but with check-ins.⁵ The difference from full extinction is real but not enormous.

What the Evidence Actually Shows

There is more research here than in most parenting topics, and it is worth reporting precisely rather than in either direction of spin.

The overall picture

An American Academy of Sleep Medicine review covering 52 treatment studies found that 94 percent reported efficacy, and over 80 percent of treated children showed clinically significant improvement maintained at 3 to 6 months.² The evidence base skews toward lower-tier study designs, with 9 studies at the highest level and 26 at level three.

The randomized trials

A trial of 156 mothers of infants aged 6 to 12 months with severe parent-reported sleep problems found the problem resolved in 53 of 76 in the intervention group against 36 of 76 controls at 2 months. Among mothers with raised depression scores at baseline, resolution was 26 of 33 against 13 of 33, and depression scores fell substantially further in the intervention group.³

Two honest caveats on that trial from the authors themselves. Neither the investigators nor the mothers were blind to group assignment, which could bias in favor of the intervention. And by 4 months the difference in sleep problems was no longer significant, which they attribute to natural improvement or families stopping the strategies.³

A second trial randomized 43 infants aged 6 to 16 months to graduated extinction, bedtime fading, or sleep education. Both active approaches produced large decreases in sleep latency. Graduated extinction specifically reduced the number of awakenings and wake after sleep onset. Salivary cortisol showed small to moderate declines in both treatment groups compared with controls, and at 12 months there were no significant differences in emotional or behavioral problems or in attachment classification.⁶

The long view

The most useful study for a parent is a five-year follow-up of 326 children whose parents received behavioral sleep techniques at 8 to 10 months. At age 6, there were no significant differences on any outcome: emotional and behavioral health, sleep, cortisol, child-parent relationship, or maternal depression. The authors concluded the techniques have no marked long-lasting effects, positive or negative.⁷

That is the strongest available refutation of lasting psychological damage. It is also a refutation of lasting benefit. The value of sleep training, on the best evidence, is short to medium term.

How Many Babies This Is Actually Relevant To

The Ferber method targets sleep-onset associations. Our State of Baby Sleep 2026 report, a survey of 74,379 parents across 112 countries, shows how widespread those associations are.

Fewer than 1 in 10 babies fall asleep on their own. Our data also found that among 7 to 9 month olds fed back to sleep, 70 percent woke three or more times a night, against 54 percent of same-age babies resettled other ways.

That is a real difference and also a modest one, and it is correlational rather than causal. It is consistent with the sleep-onset association mechanism Ferber describes, but it is not evidence that any particular family needs to intervene. Plenty of babies feed to sleep and sleep fine.

Where Both Sides Overstate

This is the section most articles skip, and it is the one that will actually help you evaluate what you read elsewhere.

Opponents overstate the harm evidence

The study most often cited as showing hidden harm involved 25 infants in a five-day inpatient sleep program, settled by program staff in an unfamiliar residential facility, with saliva sampled on two days.⁸ It has no control group, no home baseline, no follow-up, and no replication.

Most importantly, the paper does not report that infant cortisol rose from baseline. It reports that mother and infant cortisol stopped tracking together. That is a correlational claim about two people’s hormone levels, not a demonstration of infant distress. The authors themselves conceded in a published exchange that the work is early and small and not conclusive.

Proponents overstate the safety evidence

The 43-infant trial is routinely described as having proven sleep training safe. It did not. With roughly 14 infants per arm, the attachment comparison cannot detect anything short of an enormous effect.⁶ That is weak evidence of absence, not evidence of safety. The cortisol sampling was morning and afternoon rather than at the moment of separation, so it measures daily hormone tone rather than acute distress.

The fair summary of the attachment literature is two sentences, and both belong together: no study of any design has demonstrated an adverse attachment effect from behavioral sleep intervention, and no study has been adequately powered to rule one out.

On AAP endorsement

You will see sleep training described as AAP recommended. There is no AAP policy statement or clinical practice guideline on behavioral sleep interventions. The AAP publishes the primary trials in its journal, which is publication rather than endorsement. Its binding policy in this area is safe sleep. HealthyChildren does advise putting babies down drowsy but not already asleep from around 4 months.⁹

How Ferber Compares to Other Methods

One pattern in that table is worth naming out loud, because it is uncomfortable and nobody says it: the evidence gradient runs roughly opposite to the gentleness gradient. The methods involving the most crying have the most research behind them. The ones marketed as gentlest have the least.

There are two genuine exceptions, and both are worth knowing about. Bedtime fading and preventive parent education are both low-crying and both genuinely evidence-supported. Bedtime fading in particular is under-recommended relative to its evidence, and it was an active arm in the same trial that tested graduated extinction.⁶ If you want results without extinction, that is the option to look at first.

When Not to Use It

No organization publishes a formal contraindication list, but the clinical criteria and trial exclusions point clearly at these:

• Untreated reflux or pain. Pain-driven waking will not extinguish, because the crying is not a learned signal.
• Poor weight gain or a medically indicated night feed. Extinction removes night feeds.
• Suspected sleep-disordered breathing. Snoring, mouth breathing or pauses should trigger evaluation, not sleep training. Behavioral work will not fix it and delays the right workup.
• Prematurity. Use corrected age.
• Acute illness, fever, ear infection, or teething. Pause and resume once well.
• Recent major change. A move, a new sibling, a new caregiver, travel, or a return to work.
• Neurological or developmental conditions affecting arousal. These children were excluded from the trials.
• Untreated maternal depression or thoughts of self-harm. These were trial exclusions and need care in their own right.

And a legitimate non-reason to do it: if night waking is not a problem for your family, there is nothing to treat. Ferber’s own 2006 position is that co-sleeping works fine for many families.⁴

One more thing worth weighing on the other side. Persistent infant sleep problems are strongly associated with maternal depression, and in the trial data a persistent sleep problem was the variable predicting increased depression scores.³ Maternal depression is not a minor endpoint. For a family where sleep deprivation is genuinely damaging, that is a real benefit to weigh against a short period of crying.

Whatever you choose, safe sleep is a precondition rather than an add-on. See our guides to sleep regression ages and wake windows by age for the timing context.

Frequently Asked Questions