What does the research really say about cry-it-out? Not what you've been told — what the evidence shows. Honest, balanced, no agenda.
Somewhere along the way, sleep training became something you’ll eventually need to do. Not might do. Need to. If your baby isn’t sleeping through by a certain age, the advice — from books, websites, well-meaning relatives, and sometimes clinicians — converges on a single recommendation: it’s time to sleep train.
If you haven’t done it, you feel behind. If you tried and it didn’t work, you feel like a failure. If it felt wrong and you stopped, you’re told your feelings are the problem — a weakness to be restructured, not a signal to be trusted.
This article is not going to tell you what to do. It is going to give you the research, honestly presented, so that whatever you decide is an informed decision — not someone else’s recommendation, anxiety, or revenue model.
The evidence on cry-it-out is real. It is also more nuanced, more conditional, and more limited than you’ve been told.
For the commercial context behind how this advice is delivered, see The Sleep Training Industry: How Certainty Gets Sold. For permission to choose a different path, read You Don’t Have to Sleep Train Your Baby. If you’re mid-attempt and it doesn’t feel right, go to If Cry-It-Out Doesn’t Feel Right, Trust Your Instincts.
“Cry-it-out” is used colloquially to describe several distinct approaches, and the distinction matters for reading the evidence.
Unmodified extinction is the true CIO method: the parent puts the baby down and does not return until morning, regardless of how long the baby cries. Developed in the early twentieth century, most associated with Emmett Holt.
Graduated extinction — the Ferber method — involves putting the baby down and checking in at progressively increasing intervals, typically beginning at three to five minutes. The check-ins are brief and non-stimulating. Richard Ferber’s own approach has evolved substantially since his 1985 book, and the current guidance is less rigid than the popular version.
Controlled comforting / controlled crying covers variations of graduated extinction with different check-in structures. Terminology varies by country and consultancy.
Camping out / gradual retreat involves the parent remaining present and progressively increasing distance from the baby over several nights. Less distressing for most parents than extinction-based methods.
Bedtime fading adjusts the baby’s bedtime to align with when they naturally fall asleep, reducing the period of resisting sleep. Not extinction-based.
Responsive settling maintains parental presence while gradually reducing the level of active support — rocking less, patting rather than picking up, eventually moving to verbal reassurance.
Most commercially successful sleep training programs — the branded courses, apps, and consultancy packages — use variants of graduated extinction, often under proprietary names. This distinction matters: most published research examines extinction and graduated extinction specifically. The findings cannot be generalised to all approaches described as “sleep training.”
The most cited review of the sleep training literature is the American Academy of Sleep Medicine meta-analysis conducted by Mindell et al. in 2006. It reviewed 52 studies and found that extinction and graduated extinction were classified as effective in 17 out of 19 studies measuring behavioural sleep problems [1]. This is a genuine finding and deserves straightforward acknowledgement.
Sleep training, when it works, can meaningfully improve maternal mental health. Multiple studies document reductions in maternal depression following sleep training intervention. This is a real and important benefit — parental wellbeing is not a trivial outcome, and exhausted parents are not well-served by pretending otherwise.
A large-scale 2023 survey (Kahn et al., n=2,095) found no difference in quality of life or perceived attachment security between parents who used CIO methods and those who used gentle approaches [3]. This is reassuring data for parents who have used sleep training and worried about the relational consequences.
Most sleep training research measures parental reports of sleep improvement. When objective measures are used — video monitoring, actigraphy — the results are considerably less clear. Babies frequently continue waking at similar rates. What changes is not the waking — it is the signalling. A baby who isn’t calling out is not necessarily sleeping better [2].
The evidence under six months is particularly limited. Douglas and Hill (2013) conducted a systematic review of early sleep interventions (before six months) and found minimal impact on sleep outcomes, alongside increased maternal anxiety in some cases and undermining of breastfeeding relationships [4]. Most paediatric guidelines recommend waiting until at least four to six months for any formal sleep intervention.
Long-term outcomes: the most rigorous follow-up study available found no differences at five years between sleep-trained and untrained babies in emotional development, behaviour, or parent-child attachment [2]. This should be read in both directions: no evidence of lasting harm, but also no evidence of lasting benefit. The difference disappears.
Real-world failure rates are substantially higher than the published literature suggests. Macall Gordon, writing in Psychology Today (2025), reports that in real-world practice 40–50% of parents cannot complete extinction methods. Even in research settings, failure rates range from 16–50% [5]. The published success rates reflect research populations who completed the protocol. Many families don’t.
Publication bias compounds the picture. The overwhelming majority of published sleep training studies report positive outcomes. Studies reporting null or negative results are systematically less likely to be written up and accepted. The published record overstates the evidence in favour [2].
The expectation that babies should sleep independently and through the night is not biologically derived — it is a Western post-industrial cultural norm. Cross-cultural and historical research documents that in most of human history and most of the world today, co-sleeping and responsive night parenting are standard [6]. Night waking, which is developmentally normal, is treated in Western parenting culture as a problem requiring intervention.
The biological reality: parents are wired to respond to infant crying. Cortisol rises measurably when a parent hears their baby cry. Heart rate increases. In breastfeeding parents, milk lets down. This response is not a design flaw in the caregiving system. It is the caregiving system working as intended.
When sleep training researchers identified this instinctive response as a barrier to adherence, the research response was to develop techniques to help parents override it. Whittall et al. (2023) published work on “cognitive restructuring” — the mental reframing techniques used to help parents tolerate ignoring infant crying [7]. The research goal was making extinction easier to complete, not questioning whether it should be the first-line approach.
The commercial dimension is equally real. The baby sleep industry is a multi-billion-dollar sector whose financial model depends on parents believing infant sleep is a clinical problem requiring professional intervention. Night waking is reframed as a disorder. Feeding to sleep is reframed as a bad habit. The urgency is commercial as much as it is clinical. For more on this, see The Sleep Training Industry.
There is no single right answer. These are the genuine options, each presented fairly.
Formal sleep training (graduated extinction / Ferber method): when it works — and for many families it does — it can meaningfully reduce sleep fragmentation and parental exhaustion. If both parents are comfortable with the approach, the baby responds relatively quickly, and the household’s wellbeing improves as a result, this is a valid choice with a real evidence base. Not universal, not right for every family, but real.
Responsive and gentle approaches: bedtime fading, gradual retreat, and responsive settling are less studied but have emerging support. Blunden et al. (2022) found that responsive settling produced less distress and comparable sleep improvement [8]. Bedtime fading has been found equivalent to graduated extinction for bedtime resistance without the crying component. These approaches require more parental presence and often more time, but they work for families who cannot or don’t want to tolerate the crying arc of extinction.
Wait and support: for most of human history and most of the world today, this is the default approach. Help your baby to sleep in whatever way works — feeding, rocking, contact — and trust that developmental maturation will bring longer stretches over time. You will likely be more tired short-term. Long-term, the outcomes are indistinguishable from sleep-trained babies. This is not a lesser choice.
Optimise without training: many sleep difficulties respond to non-behavioural changes. Overtiredness, undertiredness, hunger, cold, noise, light, and medical issues (reflux, ear infections, tongue tie) all affect sleep. A consistently dark room, white noise, appropriate wake windows, and a reliable bedtime routine can improve sleep substantially without any behavioural intervention.
All four are valid choices. The right one depends on your baby, your family, your values, and your capacity — not on what anyone else’s baby did or what a sleep programme’s website describes as clinically necessary.
The evidence on cry-it-out is real, and you should respect it. It is also more limited, more conditional, and more contested than you have typically been told. It is not a universal solution and it is not a universal danger. It works for some families and not others. It has genuine benefits and genuine failure rates. Alternatives exist and have their own evidence.
The most important thing is not which method you choose. It is that the choice is yours — freely made, with accurate information, in a way that works for your family.
References below.
Maybe you're sitting outside the door right now. Every cell in your body is telling you to go in. It's okay to go get your baby. Your instinct is not the problem — it's your deepest wisdom.
Why does everyone tell you to sleep train? Is it because the evidence is overwhelming? Or is something else going on? The sleep training industry is worth billions — and its business model requires you to believe your baby has a problem.
Do you have to sleep train your baby? No. That's the complete answer. Here's why the question even needs answering — and what your actual options are.