Ferber Method: How It Works and What the Evidence Shows

You are reading about the Ferber method for one of two reasons. You are considering it, or you are trying to decide whether to feel guilty about considering it. Either way, here is what it actually involves and what the research shows.

What the Ferber method is

Graduated extinction, as Dr. Richard Ferber described it, follows a simple structure [1]. You put your baby down awake at bedtime, leave the room, and return at increasing intervals to offer brief reassurance without picking them up. On the first night, you might wait 3 minutes, then 5, then 10. On the second night, you start at 5 and increase from there. The intervals lengthen across nights until the baby learns to fall asleep independently.

The check-ins serve two purposes: they reassure the baby that you are present, and they reassure you that the baby is safe. The check-in is brief, about 1-2 minutes. You enter the room, speak in a calm, quiet voice (“I am here, you are okay, it is time to sleep”), and you may pat them briefly on the chest or back. You do not pick them up, you do not feed them to sleep, and you do not linger. Then you leave again.

What you say during the check-in matters less than how you say it. The tone should be calm and boring. If the check-in becomes an extended soothing session, or if your presence escalates the crying rather than briefly calming it, the check-in is doing more harm than good. Some babies find the brief appearance and disappearance of a parent more distressing than sustained absence; this is worth watching for and adjusting around.

What the evidence shows

The Ferber method is one of the most studied behavioral sleep interventions. The Mindell et al. 2006 meta-analysis examined 52 studies on behavioral treatments for pediatric sleep problems and found strong evidence that both graduated extinction and full extinction effectively reduce night wakings and bedtime resistance [2]. The evidence rating was the highest available in the review, and it applied to both methods. Gentler methods received a lower, though still positive, evidence rating.

The numbers from individual trials: in the Gradisar 2016 randomized controlled trial, graduated extinction reduced the time it took babies to fall asleep by about 13 minutes on average, and infants in the graduated extinction group woke less often than controls [4]. These are modest effect sizes, which is worth noting. The method works, but it does not produce instant results. It produces a measurable, consistent shift in the right direction.

The Hiscock et al. cluster RCT of 328 infants found that behavioral sleep intervention reduced infant sleep problems at 10 months and, importantly, reduced maternal depression scores at 4 months post-intervention [5]. This finding matters because parental mental health is often the variable that gets lost in sleep training debates. A parent who is not sleeping is a parent whose capacity for responsive caregiving during the day is diminished.

Most families see significant improvement within 3-7 nights [2]. The second or third night is often the hardest. There is a well-documented phenomenon called the “extinction burst,” where crying increases before it decreases. This is the baby testing whether the old rules still apply. It is a learning process, not a distress signal: the baby is checking whether the previous conditions (being held, fed, or rocked to sleep) are still available on demand. When they find consistently that the conditions have changed, the testing decreases. The extinction burst is unpleasant, and it is typically the point at which the method is either continued or abandoned.

The harm question

This is what you came for, probably. You want to know whether this will hurt your baby.

The best available evidence says no. The Price et al. study followed 225 families for five years after a sleep training intervention that included graduated extinction [3]. At age 6, there were no measurable differences between trained and untrained children on emotional health, behavior, sleep quality, cortisol levels, or parent-child attachment. The study used standardized measures across all domains, and the findings were straightforward: the trained and untrained groups were indistinguishable on every metric examined.

The Gradisar 2016 trial measured salivary cortisol in infants during and after graduated extinction and found no significant differences from controls at any point, including at 12-month follow-up [4]. This matters because a central concern about the method is that it elevates stress hormones. This study measured them directly and found no elevation.

One study that is frequently cited in the other direction is Middlemiss et al. (2012), which found a “cortisol asynchrony” between mothers and infants during extinction training [6]. It is worth understanding what this study actually showed: in a sample of 25 infants (no control group), infants’ cortisol remained elevated on one night even though their behavioral distress had decreased. This is a single-night measurement in a very small sample with no comparison group. It does not demonstrate harm; it demonstrates that the physiological response to a new situation takes time to resolve, which is true of many novel experiences in infancy.

Who it works best for

The method tends to work well for babies who have a predictable temperament and who respond to the check-ins with brief calming followed by renewed effort to sleep. These are babies who cry when you leave, quiet when you return, and cry again when you leave, but with decreasing intensity across the intervals. The check-ins function as intended: brief reassurance that helps the baby tolerate the next interval.

It works less well for babies whose temperament makes the check-ins stimulating rather than reassuring. Some babies become more activated by seeing and then losing a parent. Each check-in resets their arousal to peak levels rather than providing a moment of calm, and the crying escalates rather than following the expected pattern of initial increase and then decrease. For these babies, longer intervals, full extinction (no check-ins), or a gentler method entirely may be more effective.

It is not appropriate for every baby or every family. Some parents find the crying intolerable regardless of what the evidence says about outcomes. That is a legitimate reason to choose a different approach. The evidence supports multiple methods, and a gentler method done consistently will eventually produce similar results, though over a longer timeframe.

The method also requires consistency from everyone involved in bedtime. If one parent does the check-ins and another picks the baby up, or if a caregiver uses a different approach on certain nights, the inconsistency extends the learning process. Before starting, make sure everyone in the household is on the same page about how the nights will go. Have the conversation during the day, when everyone is rested enough to commit. A plan agreed to at 7 pm holds up better at 2 am than one improvised in the dark.

What to expect

Night 1-3: The hardest stretch. Crying may last 20-45 minutes at bedtime. The extinction burst, if it happens, usually peaks on night 2 or 3. You may see increased waking in the first half of the night as well. This is normal and expected.

Night 4-7: Crying at bedtime typically drops to under 10 minutes. Middle-of-night wakings begin to resolve. Some babies still wake once or twice but settle more quickly.

Week 2: Most babies are falling asleep with minimal fussing and sleeping longer stretches. Some wakings may persist if the baby is genuinely hungry; many sleep consultants distinguish between habitual wakings and hunger-based wakings at this stage. A full feed taken efficiently is likely hunger; a brief comfort suck that trails off quickly is likely habit.

If there is no improvement after 10 consistent nights, consider whether something else is going on: illness, teething, a room that is too warm, or check-ins that are stimulating rather than soothing. Also consider whether the method has been applied consistently; a single night of reverting to the old approach resets the learning process and extends the timeline.

A note for tonight

If you are trying this tonight and the crying is harder than you expected: you can stop. You can pick your baby up. You can try again tomorrow, or next week, or never. The method works, but only if you can sustain it, and “I cannot listen to this” is not a failure of character. It is information about which approach works for your family.

If you are on night 2 or 3 and the crying seems worse than night 1, that is likely the extinction burst, and it typically means the process is working. The increase in protest is a sign that the baby is noticing the change and testing whether the old rules still apply. If you can hold course through this stretch, the other side is usually significantly better.

If you’re concerned about your baby’s sleep or health, speak with your pediatrician or family doctor. The American Academy of Pediatrics has sleep safety guidance at healthychildren.org. If your own sleep deprivation is affecting your daily functioning or mental health, that is also worth raising with your doctor.