If your baby wakes frequently, needs help falling asleep or wants you close at night, sooner or later somebody will probably suggest sleep training. Sometimes it is presented as one possible option. More often, it sounds like a developmental necessity: your baby needs to learn to self-soothe, needs to learn to fall asleep independently, or will never sleep well unless you teach them.
That can create an enormous amount of pressure, particularly when you are already exhausted.
So let me start with the most important part. No, you do not have to sleep train your baby. Sleep training is one way some families try to change sleep and settling. It is not a developmental milestone, and there is no evidence that every baby needs to be taught to sleep independently in order to develop healthy sleep.
What do we actually mean by sleep training?
This is surprisingly difficult to answer because the term covers very different approaches. At one end are methods involving leaving a baby to cry for gradually increasing periods before responding. At the other are approaches where parents stay beside their child and gradually reduce the amount of help they provide. Some programmes also include changing bedtime, naps, feeding or routines.
This matters when we look at research. A study evaluating one behavioural intervention cannot automatically tell us whether every method described online as “sleep training” is effective or safe. It also matters when parents talk to each other. Two people may both say they sleep trained their baby while describing completely different experiences.
Does sleep training work?
For some families, behavioural sleep interventions can reduce parent-reported sleep problems and improve aspects of infant sleep. Randomised trials have found improvements in outcomes such as sleep onset, night waking or parental perceptions of sleep following some behavioural interventions. Parents may also report improvements in their own sleep, mood or quality of life.
That matters. Parental exhaustion matters, and I would never dismiss an intervention that a family feels has genuinely improved their life.
But “sleep training works” is still too simple a conclusion. The effects found in research are often modest, outcomes differ between studies, and many studies rely heavily on parent-reported sleep. Objective measures such as actigraphy or video can sometimes tell a more complicated story. A baby may, for example, stop signalling for a parent at every waking without necessarily stopping all physiological waking during the night. Those are not quite the same outcome.
What about the studies showing that sleep training is safe?
You may have heard that research has proved sleep training causes no harm. There are studies that provide reassurance. One well-known randomised trial compared graduated extinction and bedtime fading with sleep education and found improvements in some sleep outcomes without detecting differences in attachment or emotional and behavioural outcomes at twelve-month follow-up. Another long-term follow-up of an earlier behavioural sleep intervention found no evidence of differences in child emotional or behavioural health, parent-child relationships or attachment several years later.
These findings are important. They do not support claims that the behavioural interventions studied inevitably damage attachment or cause later emotional problems. But they also do not prove that every form of sleep training is harmless for every baby. A study can tell us what researchers measured in a particular group of families after a particular intervention. It cannot establish the complete absence of every possible effect, particularly when samples are relatively small or outcomes are difficult to measure.
I think parents deserve that distinction rather than being frightened in one direction or reassured beyond what the evidence can actually show.
Does crying during sleep training cause toxic stress?
This claim also needs care. Crying itself is not evidence that a baby is experiencing toxic stress. Babies cry for many reasons, and stress physiology is considerably more complicated than counting minutes of crying.
At the same time, I do not think we need to pretend that crying is irrelevant. Crying is communication. If your baby is distressed, it is reasonable to respond. You do not need scientific proof that responding to your crying baby is beneficial before you are allowed to pick them up.
One frequently cited study measured cortisol during a residential sleep programme and has sometimes been used to argue that babies remained physiologically distressed after they stopped crying. The study was small, had important methodological limitations and did not include the kind of control group needed to support many of the conclusions subsequently made about it.
So I would not use that study to tell you that sleep training has been proven to damage babies. But I would not use the limitations of that study to tell you that you therefore need to ignore your baby’s distress either.
Does my baby need to learn to self-soothe?
This is probably the idea I hear most often. Babies gradually develop increasingly sophisticated abilities to regulate their emotions and behaviour, but early regulation happens largely with other people. Your voice, touch, movement, feeding and presence can all be part of how your baby becomes calm enough to sleep.
That does not prevent development. As babies grow, they may naturally begin doing more themselves. One baby may suck their fingers, roll into a favourite position and fall asleep. Another may still want a cuddle. A toddler may need you beside them for a while and then suddenly begin talking to their teddy until they fall asleep.
There is enormous variation. Helping your baby fall asleep does not mean you have failed to teach them an essential skill.
But won’t they always need me if I keep helping?
No. Children change. A newborn who sleeps almost exclusively against another human body does not usually become a seventeen-year-old who requires rocking before school. The way your baby falls asleep at three months does not determine how they will fall asleep at three years, and neither determines how they will sleep as an adult.
You can also change something later.
This is important because parents are often encouraged to remove feeding, rocking or contact before it has actually become a problem, simply because it might theoretically become difficult in the future. If breastfeeding to sleep takes ten minutes and everybody is happy, you do not need to replace it with forty-five minutes of something you dislike in order to prevent a hypothetical problem six months from now.
What if I am exhausted and what we are doing is no longer sustainable?
Then that matters. Saying that you do not have to sleep train does not mean that you have to continue doing something that is making you miserable. Perhaps your baby wakes very frequently and you are barely functioning. Perhaps settling now takes hours. Perhaps feeding to sleep worked beautifully until recently and you simply do not want to do it every single time anymore.
Responsive sleep support does not mean never changing anything.
We can look at what is actually happening across the whole twenty-four hours: sleep timing, naps, feeding, development, settling, your baby’s individual sleep need and what happens during night waking. Sometimes a relatively small change makes a considerable difference.
You can also introduce new ways of settling while continuing to respond to your baby. Another caregiver can become involved. You can gradually reduce rocking, change where settling happens or move from feeding all the way to sleep towards another form of closeness if that is what you want.
Change does not have to begin with leaving your baby to cry.
And what if I choose sleep training?
I do not think parents need judgement from either direction. Some families decide that a behavioural sleep intervention is the right choice for them. If you do, I would want you to understand the method, have realistic expectations and feel able to stop or change course if it does not feel right.
I would also be cautious about anyone promising that a particular method will make every baby sleep twelve uninterrupted hours or permanently eliminate night waking.
Babies are still babies after sleep training. They become ill, grow teeth, travel, develop new skills, need reassurance and sometimes wake because they are hungry or uncomfortable. No intervention removes normal development from the equation.
You are allowed to choose what works for your family
The sleep-training debate often leaves parents feeling that they must join one side. You do not. You can believe that the evidence does not show inevitable long-term harm from the behavioural interventions that have been studied and still decide that leaving your baby to cry is not something you want to do.
You can respond to every cry and still make changes to sleep when your current situation is exhausting you. You can feed your baby to sleep for as long as it works and change it when it stops working.
And if your baby wakes frequently but you are both doing well, you do not need to fix their sleep simply because somebody else’s baby sleeps differently. Sleep training is an option. It is not an obligation.
Research
Gradisar M, Jackson K, Spurrier NJ, et al. (2016). Behavioral interventions for infant sleep problems: a randomized controlled trial. Pediatrics, 137(6), e20151486.
Price AMH, Wake M, Ukoumunne OC, Hiscock H. (2012). Five-year follow-up of harms and benefits of behavioral infant sleep intervention: randomized trial. Pediatrics, 130(4), 643–651.
Hiscock H, Bayer JK, Hampton A, Ukoumunne OC, Wake M. (2008). Long-term mother and child mental health effects of a population-based infant sleep intervention: cluster-randomized, controlled trial. Pediatrics, 122(3), e621–e627.
Middlemiss W, Granger DA, Goldberg WA, Nathans L. (2012). Asynchrony of mother-infant hypothalamic-pituitary-adrenal axis activity following extinction of infant crying responses induced during the transition to sleep. Early Human Development, 88(4), 227–232.