This piece is part of Scientific American’s column The Science of Parenting.
In countries like the U.S. and U.K., “training” a baby to sleep through the night is practically a rite of passage—one endorsed by more than six in 10 baby books, the American Academy of Pediatrics and countless parenting experts and baby brands.
To tired caregivers everywhere, sleeping through the night is a siren song on par with a baby who can change their own diaper or feed themselves. And, for some parents, severely disturbed sleep can run risks beyond mere irritability or tiredness: postpartum sleep loss has been linked to disorders from postpartum depression to psychosis. It’s important to note that the connection between sleep and mental health likely goes both ways; in women with bipolar disorder, for example, sleep loss can trigger mania—but insomnia itself is also a symptom. Still, no one would argue parents being exhausted isn’t a problem. It’s little surprise, therefore, that some consultants and companies promoting sleep training have social media followings in the millions.
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
Exact definitions of sleep training vary, but the general idea is that, by deliberately limiting your response to your restless or crying baby, you can help them fall asleep independently—and stay asleep all night. Two of the most common approaches—called “extinction” methods given their purpose to “extinguish” a baby’s signaling (i.e., crying) for a caregiver—are “controlled crying” (leaving a baby to cry on their own for set, increasing periods of time before soothing them) or “cry-it-out” (often understood to be leaving them to cry for as long as it takes for them to fall asleep). Other, “gentler,” versions include those like “camping out” or the “chair method,” where parents sit next to a crib and gradually move further away.
While they may seem modern, these methods date back to the Victorian era. Before that, most babies slept with family members, particularly mothers, and waking and tending to a baby overnight was largely unremarkable—a custom that remains the case for many cultures today. But in the West a cascade of changes brought about by the industrial revolution—including our shifting attitudes towards sleep as something that needed to happen in one consolidated block overnight, and our emphasis on children’s independence meant that, for the first time in thousands of generations, babies were expected to sleep alone and through the night.
The promotion of methods for getting babies to do so soon followed. “It is astonishing how soon some children find out the way to obtain what they want, and as all infants instinctively crave for their mother's presence, so they will certainly prefer her lap, and will cry for it at first,” a London surgeon wrote in his 1857 manual for mothers. But, he added, “if they are left to go to sleep in their cots, and allowed to find out that they do not get their way by crying, they at once become reconciled, and after a short time will go to bed even more readily in the cot than on the lap." More famously, this sentiment was popularized by the “father of pediatrics” Emmett Holt, who wrote in his 1894 manual The Care and Feeding of Children that “In the newly born infant, the cry expands the lungs.” He advised mothers that a baby “should simply be allowed to ‘cry it out.’ This often requires an hour, and in extreme cases, two or three hours. A second struggle will seldom last more than 10 or 15 minutes, and a third will rarely be necessary.” Update the language slightly, and these could be the words coming out of some of the most popular sleep-training accounts on Instagram today.
But does it work? After my daughter was born, trying to sort through how our own family might approach sleep, I deep-dived into the subject. What I found surprised me. The research is far more limited—and flawed—than I had expected. While sleep training is often sold with the premise that it helps babies sleep better, the evidence shows they sleep more or less the same as babies who are not sleep-trained.
It does indicate that sleep training often accomplishes, at least temporarily, its main goal: getting a baby to stop “signaling” for support when they wake. For some families, this can be life-changing. And no family should be blamed or shamed for making that choice.
But when it comes to whether sleep-trained babies sleep better, most of those findings, including three quarters of the studies in one frequently cited 2006 review, come from parents’ sleep diaries. If a baby has learned not to “signal,” parents are unlikely to know each time they wake. The few studies done using objective sleep measures, like video or actigraphy (monitoring movements to assess sleep-wake patterns), have found sleep training has little, if any, effect on a baby’s own sleep.
Another common claim: babies must be taught to “self-settle” to maintain good habits into childhood and beyond. But the longest-term study ever done comparing sleep-trained babies with controls found that, by age two, “sleep problems had largely resolved in both groups.” By age six, sleep-trained children showed no parent-reported difference in their sleep patterns, or on any other indicator measured, from their peers.
It’s also worth noting that, while this study is frequently cited as one of the most robust and largest ever done, it should be viewed with skepticism. This isn’t the fault of the researchers; it’s the nature of studying something this complex and personal. But it does highlight the flaws of sleep training research overall. Of the study’s 328 families in total—all mothers who said their infants had a “sleep problem”—174 families were allocated to the intervention group, where nurses who had been trained in extinction methods specifically offered sleep help at a well-child check. The “control” group received their usual well-child check, where nurses weren’t trained to offer this information. But only 100 of the intervention families took nurses up on their offer, and of those families, only 60 chose to receive in-depth information on controlled crying or camping out. (The other 40 families just wanted information on simple tips like bedtime routines, or their preference wasn’t noted down.) As is typical in long-term studies, a full third of the families were unavailable for the long-term follow-up five years later, making the final sample even smaller. But the biggest problem? Nothing, of course, kept the “control” families from asking their nurse about sleep training techniques, or pursuing it on their own—nor did anything commit the “intervention” families to completing the intervention. As a result, it’s possible that some people in the nonintervention group wound up sleep training—and vice versa.
What about risk? Experts like to assure parents that the “data” haven’t shown any long-term consequences to kids’ well-being. That’s true. It’s also true that very few studies follow up with families more than a few weeks after an intervention, and the vast majority rely on a parent’s assessment of the child’s wellbeing or bond with them—measures prone to bias.
There’s another problem, too. Parents often are told that sleep training definitely won’t negatively affect their child. But even if the research were plentiful, perfectly designed and found no higher overall risk in sleep-trained populations, that would still be an unscientific, and irresponsible, guarantee to make—particularly when it comes from those who don’t know the individual child or their psychological or medical history. Even the most touted behavioral interventions performed on consenting adults carry risk. In fact, researchers generally accept that some infants might be too vulnerable for an extinction method—which is why they often caution against doing it with babies who are under six months of age, are especially sensitive or anxious, or who have experienced trauma, like foster care.
Mainstream messaging also tends to skip over another risk: that sleep training doesn’t work for some children. One recent study found parents had to persist with even the fastest-working method, unmodified extinction (full-blown cry-it-out), for almost a month. They saw no improvement for more than a week. In another study, caregivers said they repeated training between two and five times in their baby’s first year. And more than four in 10 reported that controlled crying did not reduce the number of night wakings that they were aware of at all.
For many families, sleep training is a game changer. But the current narrative that it's all upside, for both babies and parents, doesn’t just cherry-pick data, or overlook the field’s significant limitations. It can be harmful. Mothers have told me that, when they “failed” to stick with sleep training, they thought they were letting their baby down—and felt guilt, and even anxiety and depression, as a result. Despite arguments that sleep training can help parental mental health (and there is some research to that effect), these negative experiences almost always are ignored. So are study results that find no positive mental health outcomes, such as recent findings that parents who sleep trained were no less likely to have depressive symptoms, sleep poorly or even feel tired than parents who did not—and studies that find that gentler, nonextinction alternatives may be linked to less stress and less depression in mothers.
Rather than ignoring these risks, we need to weigh them against what we do know. This includes that babies and children fare far better when their parents respond promptly, consistently and developmentally appropriately to their cues. It includes that infants, born the most immature of all primates, cannot regulate their own emotions—and can’t rationalize the difference between their crying being responded to during the day, versus at night. It includes that typically developing children all generally learn to self-settle eventually, sleep-trained or not.
And it includes the fact that alternative routes to better sleep do exist. That might be screening especially wakeful babies for conditions like low iron. For some families (and in many cultures), it may be bed-sharing, although it’s important to note that the AAP recommends against it and every family must carefully weigh the risks. It might be focusing more on what we can do to take charge of our own sleep, like going to bed when the baby does. Because how we think about our sleep affects whether we feel fatigued, mindset shifts like not using tracking apps, or reframing tough nights as temporary and not necessitating a terrible day, also can help.
And it could also include trying a gentler strategy, in which the parent never ignores a baby’s signaling (no matter what the clock says). In one such method, parents settle an infant in the usual way and always respond to their infant’s cries, but gradually withdraw their usual soothing technique—for example, rocking a baby and putting them down the moment before they fall asleep, gradually moving towards patting or stroking to sleep, then to verbal soothing—but always soothing (such as a cuddle) if a baby starts crying. One small, recent study found this approach decreased babies’ parent-perceived overnight wakings more than controlled crying, but, perhaps more importantly, resulted in less child stress (as perceived by the parent), less maternal stress and fewer symptoms of maternal depression than the controlled crying group.
Families also shouldn’t be sold false promises of better sleep based on flawed data and overconfident interpretations of its promises.
Ultimately, there’s something that no one else—not me, not your pediatrician and not a random sleep trainer shilling their services—can possibly know. And that’s whether, having weighed all of the potential risks and benefits, sleep training truly is the best choice for your family. As for any other health-related intervention, it should be one that is genuinely informed, not one anyone feels forced by circumstances, or pressured by society, to make.
As for me? After weighing up the research, its limitations, the conversations I’d had with scientists, and—perhaps most importantly—my daughter’s temperament (and my own), my energy levels and our family’s values, I decided to carry on as we were. We never tried to sleep train. Three years later, we still lay with her as she falls asleep and respond whenever she wants us—and, as the research indicated was likely, she “sleeps through” most nights, anyway.
But something less quantifiable may be more important. “I love sleeping,” she told me the other day. Is it because we’ve always supported her to sleep? Or would she have said this anyway? I’ll never know, and regardless, she is an n of 1. But I’m glad we didn’t do it any other way.
Your family may be different. And that’s okay, too.
This is an opinion and analysis article, and the views expressed by the author or authors are not necessarily those of Scientific American.