The Ferber method:
What does the evidence tell us about "cry it out" sleep training?
© 2008 - 2017 Gwen Dewar, Ph.D., all rights reserved
The Ferber method, also known as "graduated extinction," is perhaps the most well-known sleep training program for young children.
It is also one of the most controversial.
In a series of training
sessions, parents leave their children alone for strictly-timed
intervals, ignoring any protests and cries they might hear. When the method
works, children gradually accept that no one will come to their aid, and,
as a result, their behavior becomes less disruptive (Reid at al 1999).
If you're considering sleep training for your child, this article will help you decide if graduated extinction is right for you. Here I discuss the following:
- When can the experts agree about, and what key questions remain unanswered?
- How does graduated extinction work?
- What are the arguments for and against graduated extinction?
- Why hasn't the research sorted this out yet?
What are the "gentler" alternatives, and how do they measure up?
As I explain below, other methods of sleep training - methods that do not involve leaving babies alone to cry -- have been tested and judged effective. Desperate, tired parents have more than one option.
Precautions and limitations: What can the experts agree about?
Although Ferber sleep training is controversial, researchers agree on many points.
1. The Ferber method is not appropriate for young babies
Young babies need to feed at night, and their sleep-wake patterns are
still immature. For these reasons, even researchers who advocate Ferber
sleep training warn that sleep training is inappropriate for babies under 6 months old (France and Blampied 1999; Owens et al 1999).
Instead, parents can try out a number of safe, age-appropriate
tactics for improving their babies' sleep. For more information, see
these articles on infant sleep problems
infant sleep aids.
2. The Ferber method is NOT appropriate
for kids who have a conditioned fear of being left alone in their beds
or who have a conditioned vomiting response
2. The Ferber method is not appropriate for children who have a conditioned fear of being left alone or who have a conditioned vomiting response
Common sense suggests that traumatized kids should not be subjected
to graduated extinction, and advocates of Ferber sleep training agree.
If past experiences have taught your child to fear being left alone in
his crib or bed—and/or to respond to separation by vomiting—you should
consult with a behavioral psychologist about the most appropriate
approach to sleep training (France and Blampied 1999).
3. The Ferber method is not appropriate as a treatment for most child sleep problems
When parents consider sleep training, it’s usually because their children
are experiencing bedtime problems and/or disruptive night wakings. The
Ferber method seems like a possible solution. But is it?
Graduated extinction is designed for one, narrow purpose: To get children to fall asleep without parental soothing. Judged on this basis, the
method is effective (see "Arguments in favor of the Ferber
But the Ferber method is NOT designed to treat most of the sleep
problems that cause bedtime battles and night wakings. For instance, the
Ferber method doesn’t address
If one of these conditions is responsible for your child's
bedtime problems or night wakings, it's important to find appropriate
treatment. Otherwise, you risk worsening your child’s psychological
problems or ignoring potentially dangerous medical conditions.
On the positive side, you might find that treating these
conditions renders sleep training unnecessary. For instance, if your
child resists bedtime because he isn't sleepy, treating a circadian
rhythm disorder or adjusting his bedtime may solve the problem
For help identifying—and treating—your child's sleep troubles, see these articles on
4. The Ferber method does not teach children how to fall asleep
As Richard Ferber himself acknowledges, graduated extinction doesn't
teach children how to fall asleep on their own (Ferber 2006). Children are
simply denied access to their parents, and left to work it out for
If you really want to help your child fall asleep, try this alternative to the Ferber method—a sleep training program known as "positive routines with faded bedtime."
Unlike Ferber sleep training, the "positive routines" program will teach
your child how to relax and prepare for bed (Adams and Rickert 1989).
You can also help your child fall sleep by following these practices:
- Dim the lights in the evening
- Avoid exercise and stimulating activities before bedtime
- Avoid stimulants (like caffeine) and hard-to-digest foods before bedtime
Help your child cope with nighttime fears
Don't intervene the moment you hear a sound. As noted in my articles about newborns and older infants,
babies often sigh, jerk, or vocalize during sleep. You don't want to
wake a sleeping baby by mistake, and even if your baby is awakening, she
may go back to sleep spontaneously if you give her the chance.
5. Studies leave important questions unanswered
Researchers agree that babies under the age of 6 months
should not undergo graduated extinction training. But what about older babies?
Are there any risks or long-term side-effects?
Advocates of graduated extinction training note that no studies yet
have demonstrated that the Ferber method harms children over 6 months old. But
the truth is we lack well-designed, controlled studies to resolve these
In some cases, the problem is that no research exists.
instance, studies of human infants
confirm that crying is physiologically stressful—increasing a baby's blood
pressure, heart rate, and cortisol levels (Levesque et al 2000; Luddington-Hoe
et al 2002). Do the intensified crying bouts—the so-called "extinction
bursts"—associated with the Ferber method put babies at any immediate risk? No one
yet has tested this hypothesis.
In other cases, studies have been published, but these
studies aren't designed in ways that permit us to answer our questions. To see what I mean, consider the following.
Testing graduated extinction: How to design a study that addresses key concerns
A well-designed experimental study isolates those variables
that concern us, and contrasts them with a control condition that doesn't
include those variables.
In this case, one of the key areas for concern is
leaving a crying infant in a setting where the baby can't see, hear, or
otherwise detect the presence of a caregiver.
As noted below, this feature of
the Ferber method stands in stark contrast with the cross-cultural and
evolutionary norm for infant care, which is for babies to fall asleep in the
comforting presence of caregivers.
Does this divergence from the species norm change the way
To find out, we need an experiment that contrasts infant
isolation with its natural counterpart: babies going to bed with the awareness
that their caregivers are present (i.e., in the same room) and responsive
(i.e., ready to attend, rather than ignore, infants that become
We also need to take care that other factors are held
constant. This includes baseline characteristics, like stressors in the home,
but it also includes changes introduced during the experiment.
parents in one group receive special instruction in how to time feedings
optimal sleep, the same training should be given to parents in the other
condition. Otherwise, we won't be able to tease apart causation. We
which aspects of the experimental manipulation (learning about timed
feedings, or practicing the Ferber Method) affected the outcome.
Finally, we need to address crucial aspects of quality control:
- Did all parents in the treatment condition follow
the same procedures?
- If some families dropped out of the study, do we know how
this might have biased the results?
- Have the researchers devised objective
measures for determining long-term outcomes (like the development of emotional
regulation, personality, and attachment security)?
- How large is the study, and
what kind of families participated?
And before we reach any
broad conclusions, we need to consider the specific characteristics of
the study population. Can we generalize the results to other groups --
including people living in different cultures?
It's a lot to think about, but it's necessary if we want to address
questions relevant to the debate. Unfortunately, as I note below, no
study yet has met the first criterion (contrasting graduated extinction
with species-normal nighttime care), let alone all of them. So despite
decades of research, we still know surprisingly little about how the key
features of graduated extinction might influence a child's behavior,
development, and family relationships.
How the Ferber method works
The Ferber method is actually a variant of a sleep training program called "extinction."
Extinction sleep training is based on the assumption that
children have sleep problems because they have learned to depend on
parental soothing to put them to sleep. Whenever kids are denied this
parental soothing, they may stall, plead, cry, or throw tantrums to
delay bedtime. Because they can't fall asleep by themselves, kids will
also be disruptive if they awaken during the night.
Parents who "give in" to their children’s demands for attention
are reinforcing the problem behaviors. So the solution is for parents to
put their children to bed (while they are still awake) and then leave
them alone. If children cry, parents are instructed to ignore it. Parents
aren't supposed to check on the child again unless it seems absolutely
necessary (Owens et al 1999).
The Ferber method departs from extinction training in one key
respect: It permits parents to check on their children—but only briefly
and according to a strict schedule. On the first night of training,
parents put their child to bed and then stay away for 3 minutes. After a
brief check (during which the parents take care not to pick up or hold
the child) the parents leave again—this time for 5 minutes.
Subsequently, parents wait 10 minutes between visits until the child
finally falls asleep.
For each night that follows, parents gradually increase the time
between checks. For instance, on the second night, parents might wait 5
minutes before the first visit, 10 minutes before the second, and 12
minutes before all subsequent visits. On subsequent nights, these
intervals might stretch to 20 minutes or more.
Some parents who try Ferber sleep training see improvements
within a few days. In studies testing graduated extinction, parents may
complete training within 4 weeks (e.g., Reid et al 1999).
But some parents—disturbed by the notion of ignoring their
children, and concerned about the potentially harmful effects of
training—drop out before they see any improvements in their children's
behavior. Babies subjected to the Ferber method may become much more
distressed during training than they were before. These so-called "extinction bursts"--which include more frequent and intense crying,
protests, and tantrums—persuade some parents to give up.
In addition, training may fail if parents are inconsistent and
periodically "give in" to their children's pleas. Ferber advises parents
to stick to the routine, even if the child becomes so upset that he
vomits. In this event, parents should clean up the mess quickly and then
leave the room and continue training (Ferber 2006).
Arguments in favor of the Ferber method
According to scientific studies, extinction sleep training--including
Ferber sleep training--is associated with the following positive
outcomes (Mindell et al 2006):
- Children who complete training are less likely to throw bedtime tantrums
- Children who complete training are more likely to settle down at night within ten minutes
- Children who complete training are less likely to awaken their parents during the night
- Parents who complete training report improvements in their own stress levels, mood, and interactions with their children
In addition to these positive results, parents have reported
improvements in their childrens' daytime behavior, perhaps because sleep
training "graduates" were getting more sleep at night (Mindell et al
For parents enduring nights of sleeplessness and emotional
turmoil, these outcomes are extremely important. Advocates of graduated extinction out that parents who are sleep-deprived are at higher risk
for depression and marital conflict (Mindell et al 2006). Such parents
may develop negative feeling toward their children and the parent-child
relationship suffers as a result. If parents can stop or reduce their
children's disruptive nighttime behavior, the whole family will benefit.
But the Ferber method isn't the only sleep training program that delivers these benefits.
As noted below,
alternatives to the Ferber method -- training programs that don't involve leaving children alone to "cry it out"-- have similar track records (e.g., Skulladotir et al 2003; Matthey
and Črnčec. 2012). For this reason, it's important to consider the potential costs of graduated extinction.
Arguments against the Ferber method
Although studies show that extinction sleep training can be very effective in eliminating bedtime protests and stalling tactics, many people worry about potential side effects.
Leaving children alone to cry seems to violate our deepest instincts, and no wonder.
For most of human history, our ancestors ' biggest sleep problem was almost certainly the avoidance of predators. Like modern-day hunter-gatherers, our ancestors slept communally and shared "watch" duties (Worthman and Melby 2002). Children snuggled up to their parents and siblings. If children cried out, it was important to soothe them quickly.
Moreover, it was important for babies to forge close, personal ties to their caregivers. They required care and feeding for many years before they could survive on their own. In a world where 40% of children died before their 15th birthdays (Kaplan et al 2000), success depended on having somebody looking out for you -- somebody who understood your needs and was committed to meeting them. No wavering. No neglect.
So ignoring an infant's distress was a sign that something was wrong. And leaving a child alone would have constituted abandonment, if not attempted infanticide (Hrdy 2009). It therefore shouldn't surprise us if our ancestors evolved emotional and behavioral responses to keep parents and children close.
Our evolutionary past has left its stamp in our brains. When young children are left alone at night, they may experience one of the most primal and powerful stressors known to young animals--separation anxiety (Panksepp 2000). Separation anxiety is a panic response arising from a primitive part of brain that also processes information about physical pain (Panksepp 2000).
What happens when we ignore this response? And what happens when parents act as if they are insensitive to their children's distress? How does the Ferber method affect the stress response system? The development of attachment relationships, emotional skills, and personality traits?
In this context, it's important to keep in mind that we can't assume that everything is all right because a baby doesn't cry.
In a recent experiment on babies in a hospital nursery, researchers recorded brain EEGs while infants experienced different forms of stimulation, including a painful heel stick.
The researchers discovered there was no association between infant crying and evidence of pain signalling in the brain. Some babies cried when they were in pain, others did not. Some babies cried more intensely when they were in pain, others did not (Maitre et al 2017).
The absence of crying wasn't a reliable indicator of a baby's distress. Might the same be true among babies undergoing "cry it out" sleep training? When babies stop crying, it could indicate that they're no longer distressed. Or it might merely indicate that they have given up trying to signal a caregiver who doesn't respond.
These questions concern a broad array of parents, pediatricians, researchers, and therapists (e.g., Sears and Sears 1996; Commons and Miller 1998; Sunderland 2006). And, as noted above, we need more research to answer these questions. But why is that?
Why the research hasn't yet sorted this out
If you've followed stories in the
popular media, you might have gotten the impression that the controversy
is over. Haven't studies demonstrated that graduated extinction is a
safe, cost-free way to solve a child's sleep problems?
The answer is no, and the reason is straightforward: Currently we lack studies that meet our design criteria.
For example, one study -- highly touted in the media
-- tracked outcomes for more than 250 Australian infants over a period of 5 years (Hiscock et
al 2007; Price at al 2012).
That sounds very promising, but there is a problem.
This study didn't actually test the effects of
the Ferber method.
Parents participating in the intervention received multiple forms of help and advice:
They were taught about baby sleep patterns.
- They were instructed to improve infant sleep by tweaking the timing
- They were advised to put babies to bed while they were still awake.
- They were assigned to nurse who coached them on handling sleep problems.
Then parents were encouraged to choose
one of two sleep training programs -- graduated extinction, and an alternative
approach that didn't involve leaving babies alone.
Thus, the treatment condition
amounted to a mixed bag. Parents got education, support, and coaching in a number of tactics, not just the Ferber method.
What about the control group? Parents assigned to the control condition weren't
encouraged by an assigned nurse to select a sleep training program. But parents
in the control condition were free to solicit advice from their regular medical
providers, and they were free to adopt any nighttime strategies they wished.
Did some parents in the control condition attempt graduated extinction?
We don't know
because the researchers didn't collect information about this.
But we know the parents were potentially motivated, and had the opportunity.
All the parents who participated in this study were selected because they had complaints about their infants' sleep habits. And control-group parents
didn't just seek advice from their regular
medical providers. Almost two-thirds of them said they had received help
advice from outside sources.
It wouldn't be surprising, then, if some
the control group received Ferber-style sleep training.
So this study doesn't permit us to evaluate the long-term effects of
The treatment condition includes several different
strategies lumped together, including many tactics that don't involve leaving
babies alone. Moreover, babies in the control group probably experienced some
of the same training conditions that babies in the treatment group
experienced. What this study really
tested was the effects of offering parents a package of education and support.
What about more recent research?
Another group of investigators in Australia have applied a tighter focus to sleep training research (Grandisar et al 2016). They assigned some families to use graduated extinction, and contrasted the results with those of families assigned either to (1) a control condition, or (2) to a group practicing an alternative sleep strategy.
But once again, a lack of information prevents us from making inferences.
For parents in the control group, researchers didn't collect data on what techniques parents used to get their babies to sleep. Nor did they ask if babies slept in the same room with their parents, or somewhere else (Grandisar personal communication).
Likewise, for parents assigned to use the alternative sleep strategy, researchers didn't collect data on sleep location for babies.
As a result, one of the variables of greatest concern to critics -- whether or not babies were left to cry alone -- was left uncontrolled and untested.
In addition, this study suffers from limitations related to small sample size and missing data. No more than 15 families were assigned to each condition, and many families failed to participate at certain points in the study.
For example, at the 3-month data collection time point, only about half the families checked in. The researchers filled in missing data with their own estimates (Grandisar, personal communication).
Missing data can have a big impact, particularly when we're dealing with a small sample size to begin with. You can appreciate this if you consider a test the researchers conducted at the end of the study, 12 months after the start.
Only 7 out of 13 (54%) of families in the graduated extinction group had babies that were scored as securely attached to their parents. By
contrast, babies in the control group fared a bit better: 5 out of 8 babies
(63%) were scored as securely attached.
That might look bad for graduated extinction, but notice that fewer families in the control group actually participated in the test. If nobody had opted out of the attachment test, would the results have looked different?
So this research doesn't target key questions about graduated extinction, and it's limited by small sample size and missing data. As I write this in the spring of 2017, we are still waiting for studies that can provide us with strong inferences about the long-term effects of leaving babies alone and deliberately ignoring their cries.
The Ferber method compared: How does it measure up against the alternatives?
The Ferber method is not the only option for parents interested in sleep training. Two alternatives include
- "Positive Routines with Faded Bedtime," and
- "Extinction with Parental Presence," (in which parents remain in the same room with children until they fall asleep)
The "positive routines" method teaches children how to fall asleep
by pacing them through a series of relaxing bedtime rituals--but only
when the child is already showing signs of drowsiness.
parental presence" is a variant of graduated extinction that doesn't
involve leaving children alone.
(For a step-by-step account of these methods, see my article on gentle sleep training.)
How does the Ferber method stack up against these two alternatives?
In the only controlled, randomized scientific study to compare graduated extinction and "positive routines" head-to-head, there were no significant differences in
treatment outcomes for kids (Adams and Rickert 1989).
Similarly, an experiment pitting graduated extinction against "extinction with parental presence" found no difference in treatment outcomes (Matthey and Črnčec 2012).
According to the American Academy of Sleep Medicine, these three
methods "appear to be on a level playing ground" (Mindell et al 2006:
The Ferber method can be an effective way to reduce nocturnal crying,
protests, and requests for parental soothing. But it's not the only
effective method available.
Before you choose any sleep training program, review these points:
- Make sure that the sleep training program is
- Screen and address your child's specific sleep problem.
- Consider your child's individual temperament and personality.
- If your child has a conditioned fear of being alone or shows signs of conditioned vomiting, consult a behavioral therapist.
- Before training begins, make sure that all participating adults
understand the procedure. Regardless of the method you choose, success
depends on being consistent.
More information about baby sleep
For help with infant sleep problems, see these evidence-based pages.
References: The Ferber methods and its alternatives
Adams LA and Rickert VI. 1989. Reducing bedtime tantrums: Comparison
between positive bedtime routines and graduated extinction. Pediatrics
Commons ML and Miller PM 1998. Emotional learning in infants: A
cross-cultural examination. Paper presented at the American Association
for the Advancement of Science, Philadelphia, PA, February, 1998.
B. 2004. Treatment of sleep problems in families with young children:
effects of treatment on family well-being. Acta Paediatr ;93:126-134.
Ferber R. 2006. Solving your child’s sleep problems: New, revised, and expanded edition. New York: Fireside.
France KG. 1992. Behavior characteristics and security in sleep
disturbed infants treated with extinction. J Pediat Psychol 17: 467-475.
France KG and Blampied NM. 1999. Infant sleep disturbance:
Description of a problem behaviour process. Sleep Medicine Reviews 3(4):
Grandisar M, Jackson K, Spurrier NJ, Gibson J, Whitham J, Sved Williams A, Dolby R, Kennaway DJ. 2016. Behavioral Interventions for Infant Sleep Problems: A Randomized Controlled Trial. Pediatrics 137(6).
Hall WA, Hutton E, Brant RF, Collet JP, Gregg K, Saunders R, Ipsiroglu O, Gafni A, Triolet K, Tse L, Bhagat R, and Wooldridge J. 2015. A randomized controlled trial of an intervention for infants' behavioral sleep problems. BMC Pediatr. 15(1):181.
Hrdy SB. 2009. Mother nature: Maternal instincts and how they shape the human species. NY: Ballantine Books.
Kaplan H, Hill H, Lancaster J, Hurtado AM. 2000. A Theory of Human Life
History Evolution: Diet, Intelligence, and Longevity. Evolutionary
Anthropology 9(4) 156-185.
Konner M. 2005. Hunter gatherer infancy and childhood: The !Kung and others.In:BS Hewlett and ME Lamb (eds): Hunter-gatherer childhoods: Evolutionary, developmental and cultural perspectives. New Brunswick, US: Aldine.
Levesque BM, Pollack P, Griffin BE and Nielsen HC. 2000. Pulse
oximetry: What’s normal in the newborn nursery? Pediatric pulmonology
Luddington-Hoe SM, Cong X, and Hashemi F. 2002. Infant crying:
nature, physiologic consequences, and select interventions. Neonatal
Network 21(2): 29-36.
Maitre NL, Stark AR, McCoy Menser CC, Chorna OD, France DJ,
Key AF, Wilkens K, Moore-Clingenpeel M, Wilkes DM, Bruehl S. 2017. Cry presence
and amplitude do not reflect cortical processing of painful stimuli in newborns
with distinct responses to touch or cold. Arch Dis Child Fetal Neonatal Ed.
2017 May 12. pii: fetalneonatal-2016-312279. doi: 10.1136/archdischild-2016-312279.
[Epub ahead of print]
and Črnčec. 2012. Comparison of two strategies to improve infant sleep
problems, and associated impacts on maternal experience, mood and infant
emotional health: a single case replication design study. Early Hum
Mindell JA, Kuhn B, Lewin DS, Meltzer LJ, Sadeh A and the
American Academy of Sleep Medicine. 2006. Behavioral treatment of
bedtime problems and night wakings in infants and young children. Sleep
Owens JL, France KG, and Wiggs L. 1999. Behavioural and
cognitive-behavioural interventions for sleep disorders in infants and
children: A review. Sleep Medicine Reviews 3(4): 281-302.
Panksepp J. 2001. Affective Neuroscience: The Foundations of
Human and Animal Emotions. New York: Oxford University Press.
Reid MJ, Walter AB, and O’Leary SG. Treatment of young children’s
bedtime refusal and nighttime wakings: A comparison of “standard" and
graduated ignoring procedures. Journal of Abnormal Child Psychology 27:
Price AMH, Wake M, Epi GD, Ukoumunne OC, Hiscock H, Epi GD. 2012.
Five-Year Follow-up of Harms and Benefits of Behavioral Infant Sleep
Intervention: Randomized Trial. Pediatrics 130(4):643-51.(doi:
Sadeh A. 1994. Assessment of intervention for infant night
waking: Parental reports and activity-based home monitoring. J Consult
Clin Psychol 62(1):63-8.
Sears W and Sears M. 1996. The fussy baby book: Parenting your
high-need child from birth to age five. New York: Little, Brown and
Skuladottir A, Thome M, and Ramel A. 2005. Improving day and
night sleep problems in infants by changing day time sleep rhythm: A
single group before and after study. Int J Nurs Stud. 42(8): 843-850.
Sunderland M. 2006. The Science of Parenting. New York: DK Publishing.
C.M. Worthman and M. Melby. 2002. Toward a comparative
developmental ecology of human sleep. In: Adolescent Sleep Patterns:
Biological, Social, and Psychological Influences, M.A. Carskadon, ed.
New York: Cambridge University Press, pp. 69-117.
Was this article helpful? If so, try browsing through the parenting science index
or use the search button at the top of the left hand column to find something specific.
Content last modified 7/17
image of baby in crib by Joan Campderrós-i-Canas/flickr
image of father comforting baby by Andres-Nieto-Porras / flickr
image of mother comforting baby by tostadophoto.com / flickr