What the research says about behavioural sleep approaches

1. Stress hormones

What was done

Forty-three infants aged six to sixteen months, randomly assigned to graduated extinction, bedtime fading, or a control group receiving sleep education only. Saliva samples taken morning and afternoon before treatment, then at one week, one month, three months and twelve months.

What was found

Cortisol declined. Small to moderate declines in both treatment groups compared with controls, with levels staying within normal limits throughout. No sign of the sustained elevation the concern predicts, and no sign of dysregulation of the stress system.

What was not measured, and it matters

No samples were taken during the actual nights of crying. Real-time measurement in the moment was not done, and the authors say so themselves.

So a short-lived cortisol spike while a baby is crying at eight o'clock on night one cannot be ruled out. What the follow-up did not find was any lasting elevation. Most parents already assume their baby is stressed in the moment, because they can hear it. What this research addresses is whether it leaves a mark.

2. Attachment

The same infants were assessed twelve months after the intervention using the Strange Situation Procedure, the standard laboratory assessment of infant-parent attachment, in use since the 1970s. Sessions were videotaped and scored by an assessor who did not know which group each family had been in.

No significant difference in secure versus insecure attachment between the groups. The specific fear, that leaving a baby to protest damages the bond, was tested with the best available measure and was not supported.

3. Behaviour and mental health at twelve months

Emotional and behavioural problems assessed twelve months after the intervention using the Child Behavior Checklist. No significant differences between groups on internalising problems, externalising problems, or total problems.

4. Five years later

What was done

A five-year follow-up of a randomised trial. Three hundred and twenty-six children whose parents had reported sleep problems at seven months, 173 of them randomised to receive the behavioural intervention. At six years of age researchers assessed child mental health, sleep, quality of life, general health, the closeness and conflict in the child-parent relationship, disinhibited attachment, and maternal mental health and parenting style. They also collected salivary cortisol from the children as an objective stress measure.

What was found

No meaningful differences on anything. Emotional behaviour p=.8. Conduct behaviour p=.6. Sleep problems 9% versus 7%. Child-parent closeness p=.1 and conflict p=.4. Global relationship p=.9. Disinhibited attachment p=.3. Maternal depression, anxiety and stress p=.9.

The cortisol result is worth pausing on. Five years after the intervention, 28.8% of the intervention children showed an abnormal diurnal cortisol profile against 21.7% of controls, p=.4. That is an objective physiological stress measure taken half a decade later, and it found nothing.

And the authors go further than most people quoting them do. Their conclusion is that these techniques have no marked long-lasting effects in either direction. The benefits fade too, somewhere beyond two to three years. What they deliver is a real short-to-medium-term improvement rather than a permanent change in your child.

"No difference detected" is not the same as "proof that no difference could exist." These studies looked at specific outcomes with specific measures over specific timeframes, and within those limits they found nothing. That is meaningful. It is not the same as certainty.

5. The argument that has not been settled

This is the most compelling objection, and it deserves a straight answer rather than a dismissal.

Movement monitors worn by infants during clinical trials showed something intriguing. While parents' sleep diaries reported dramatic improvements, the monitors showed that babies were still stirring at night. They had simply stopped calling out.

Critics interpret this as learned helplessness, arguing that the baby has given up and stopped bothering to call. Paediatric researchers interpret the exact same data differently: the baby is waking briefly, as all humans do, and returning to sleep independently without needing parental intervention, which is the definition of self-soothing.

Both theories fit the movement data. What separates them is every other clinical measure. If infants were experiencing chronic withdrawal, you would expect that distress to appear somewhere in the saliva cortisol, in the attachment assessments, or in behavioural evaluations at twelve months and six years. It does not appear in any of them.

While not absolute proof, the weight of clinical evidence supports the self-soothing explanation. You should know both interpretations exist and judge them for yourself.

6. What this does for parents

The most profound, measurable effects in this body of research are actually on the parents.

In a randomised trial of 156 mothers of infants aged six to twelve months with persistent sleep challenges, half received behavioural sleep guidance while half received written information about normal sleep. At two months, sleep problems had resolved for 70% of the intervention group, compared with 47% of the control group.

Among mothers who met clinical criteria for postnatal depression at the start, the difference was even more striking: 79% of their babies' sleep issues resolved, compared to just 39% in the control group. Maternal depression scores fell significantly further, maternal sleep quality improved, and families required fewer outside healthcare visits.

When viewed alongside the six-year child outcomes, the clinical reality is clear: your child's emotional development and bond with you remain secure either way, which means the true, immediate beneficiary is often the parents. Sleep deprivation is not a minor inconvenience, maternal mental health is vital, and taking proactive steps to protect it is entirely legitimate.

7. The first nights are the hardest, and that is measurable

A study of 91 infants and 182 parents measured parental tolerance for crying before starting an approach and again one month later. Researchers assessed how parents interpreted night waking and how quickly they felt compelled to intervene.

Over the course of the month, parents became significantly better at distinguishing between brief, transitional protests and genuine distress. Their anxiety surrounding night wakings dropped sharply, particularly among those using structured interval settling.

When parents are told that the first few nights are the hardest, that is not just empty encouragement. It reflects a measurable shift in parental confidence and emotional regulation.

8. One method does not fit every baby

A landmark clinical trial compared two approaches directly across 91 infants aged nine to eighteen months, randomised between interval checking and a parent-present approach where the parent stays in the room and withdraws gradually.

While both approaches improved sleep across the entire cohort, the results diverged significantly based on separation anxiety. Infants with high separation anxiety who received interval checking showed almost no reduction in night waking during the first month (p=0.92). Those same high-anxiety infants improved clearly and quickly when a parent remained in the room. Lower-anxiety infants progressed well on either approach.

The researchers concluded that gentler, parent-present approaches should be favoured for infants showing high separation anxiety. Handing every tired parent the exact same interval-checking method is a clinical mistake, and it explains why so many families struggle.

9. Where we agree with the critics

Not every approach suits every baby

The research is clear: forcing a high-distress infant through rigid intervals without considering their temperament is ineffective. Matching the method to the child is essential.

Under six months, the critics are correct

Behavioural sleep interventions in the first six months are not supported by clinical evidence. A meta-analysis of nine randomised trials covering 1,656 mother-infant pairs showed no reduction in infant night wakings before six months. Circadian rhythms only begin maturing between ten and twelve weeks, and night feeding is often physiologically necessary.

Some evidence bases are smaller than claimed

Certain widely recommended techniques rest on single trials with fewer than twenty infants. We identify which methods are backed by large randomised trials and which rely on smaller pilot studies, rather than pretending all advice is equally robust.

10. Deciding not to intervene

Choosing not to use any settling technique is a completely valid choice.

Night waking in healthy infants is normal. Between seven and eleven months, healthy babies wake an average of once per night, with zero to three wakings sitting comfortably within the normal clinical range. If your family is managing well, there is no medical requirement to change anything.

The purpose of this evidence summary is not to push you into an approach you are uncomfortable with, but to ensure that whatever you choose, your decision is grounded in real data rather than guilt or internet mythology.

What is not on this page

You will not find promotional testimonials, dramatic before-and-after stories, or claims about our specific commercial programme here.

This page is dedicated exclusively to the published clinical literature on infant sleep science. Blending marketing claims with scientific evidence compromises the objectivity you came here for.